Journal of Petrology Advance Access originally published online on November 8, 2006
Journal of Petrology 2007 48(2):303-325; doi:10.1093/petrology/egl062
© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@ oxfordjournals.org
The Graveyard Point Intrusion: an Example of Extreme Differentiation of Snake River Plain Basalt in a Shallow Crustal Pluton
Craig M. White*
Geosciences Department, Boise State University, Boise, ID 83725 USA
RECEIVED
AUGUST 29, 2005;
ACCEPTED
SEPTEMBER 28, 2006
 |
ABSTRACT
|
|---|
The Graveyard Point intrusion is the only known example of a
well-exposed differentiated mafic pluton associated with the
late Miocene–Pleistocene magmatism of the western Snake
River Plain (SRP). It is exposed in a 6 km by 4 km area adjacent
to the Oregon–Idaho border, and exposures range in thickness
from 20 to 160 m. The thicker parts of the intrusion are strongly
differentiated and contain a 25–60 m thick section of
well-laminated cumulus-textured gabbros that grade upward into
pegmatoidal ferrogabbro. Evolved liquids formed sheets of Fe-rich
siliceous granophyre. At least two injections of magma are indicated
by abrupt discontinuities in the rock and mineral compositions,
and by the lack of mass balance between the bulk intrusion and
its chilled borders. The laminated gabbros are interpreted to
have formed from a tongue of augite and plagioclase crystals
that were carried in with the second pulse of magma. Following
the final emplacement of the intrusion, in situ differentiation
proceeded through a two-stage process: the ferrogabbros are
explained as interstitial liquids forced out of the crystal
mush by compaction, and the siliceous granophyres are interpreted
to be residual liquids that migrated out of the partly crystallized
ferrogabbros in response to the exsolution of volatiles. Because
the geochemical trend inferred for the mafic to intermediate
composition liquids in the Graveyard Point intrusion is similar
to the trend for many western Snake River Plain lavas, the pluton
may be a good model for shallow sub-volcanic magma chambers
elsewhere in the SRP. However, some western SRP lavas contain
anomalously high concentrations of P
2O
5 , which are best explained
by mixing within the active crustal mush column or with partial
melts of previously formed differentiated mafic intrusions.
KEY WORDS: Snake River Plain; mafic intrusions; tholeiitic; sill; granophyre
|
INTRODUCTION
|
|---|
Recent papers by Marsh (2000
, 2004
) emphasized that magmas evolve
in response to combined physical and chemical processes that
occur over a wide range of scales in spatially integrated intrusive
complexes referred to as ‘magmatic mush columns’.
However, in most young volcanic provinces the intrusive parts
of this system are hidden and their petrological features can
be only inferred from petrographic, chemical and isotopic studies
of lavas and tephra. The Snake River Plain (SRP) of Idaho and
eastern Oregon is one such young province where extensive complexes
of crustal level, possibly interconnected, mafic intrusions
have been indicated by geophysical evidence, but remain largely
unknown because they have not been exposed by erosion (Mabey,
1982
; Sparlin
et al., 1982
; Peng & Humphreys, 1998
). The
SRP contains a compositionally diverse suite of mainly tholeiitic
lavas, many of which appear to have evolved by fractional crystallization
over a range of crustal depths (e.g. Leeman & Vitaliano,
1976
; Leeman, 1982
; Reid, 1995
; Geist
et al., 2002
). For this
reason, the concept of a magmatic mush column may be particularly
applicable to this province, and knowing the nature of the plutonic
system becomes a critical part of understanding the magmatic
system. To that end, this paper presents results of the first
detailed study of a western SRP intrusive complex and suggests
a physical model for the evolution of magmas in the upper part
of the mush column beneath this well-known volcanic province.
|
GEOLOGICAL SETTING
|
|---|
The western SRP is a NW-trending intracontinental rift basin
about 70 km wide and 300 km long (
Fig. 1a). The boundary faults
are parallel to major structural trends in the Pacific Northwest,
but perpendicular to the assumed NE-trending track of the Yellowstone
hotspot (Pierce & Morgan, 1992
; Smith & Braile, 1994
).
Although there is no consensus on the origin of the western
SRP, many workers have suggested that the 17–14 Ma episode
of voluminous magmatism attributed to the Yellowstone hotspot
either directly or indirectly initiated the formation of the
western SRP (e.g. Geist & Richards, 1993
; Glen & Ponce,
2002
; Shervais
et al., 2002
). The rifting and subsidence that
appears to have begun about 10 Ma is commonly attributed to
regional extension that occurred throughout the northern Basin
and Range Province (e.g. Hooper
et al., 2002
). Basaltic magmatism
in the western SRP began with the onset of extension and has
produced a diverse suite of tholeiitic lavas, many of which
are strongly enriched in iron (FeO* >14%) (Bonnichsen &
Godchaux, 2002
). More than 2 km of sediments fill the deepest
parts of the western SRP basin in southwestern Idaho (Wood,
1994
), but the basin shallows rapidly to the west and dies out
in eastern Oregon where it abuts the older, north–south-trending
Oregon–Idaho graben (Cummings
et al., 2000
). Erosion has
exposed the Graveyard Point intrusion because it was emplaced
near the western margin of the plain where the least amount
of subsidence has occurred.
View larger version (39K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 1. (a) Map showing the location of the Graveyard Point intrusion and the general outline of the western Snake River Plain in southwestern Idaho and eastern Oregon. Inset illustrates the track of the Yellowstone hotspot from southeastern Oregon to its present location beneath Yellowstone National Park (Y). (b) Simplified geological map of the eastern half of the Graveyard Point intrusion [modified from (Ferns, 1989)], showing the locations of the three measured sections. Circles indicate the locations of samples from the chilled borders of the intrusion; the numbers refer to analyses given in Table 1.
|
|
|
GEOLOGY OF THE INTRUSION
|
|---|
Age and field relations
Mafic intrusive rocks forming the Graveyard Point complex are
exposed in a series of discontinuous outcrops within an area
of
c. 6 km by 4 km. The magmas were emplaced into middle Miocene
fluvial and lacustrine sediments and silicic pyroclastic rocks
originally mapped by (Kittleman
et al., 1967
) as part of the
Sucker Creek Formation. (Ferns, 1989
) cited a K–Ar whole-rock
date of 6·7 ± 0·4 Ma for diabase from the
lower part of the main intrusion, a value that is only slightly
younger than
40Ar–
39Ar ages of basalts of similar composition
erupted along the southern margin of the western SRP in Idaho
(White
et al., 2002
). Mafic rocks of the Graveyard Point complex
occur in three types of exposures: (1) 1–10 m wide dikes
of olivine diabase; (2) 20–30 m thick sills, also composed
of olivine diabase; (3) 100–160 m thick, irregularly shaped
but generally sheet-like intrusions composed of olivine diabase,
gabbroic cumulates, ferrogabbro and granophyre. Chilled borders
in all of the floored intrusions are chemically and petrographically
similar to one another and to the fine-grained diabase in the
dikes. The field relations are consistent with a structural
model in which all exposures of the sill-like bodies are part
of a single, wedge-shaped intrusion that was offset by normal
faults.
Igneous stratigraphy
The thickest and most complete sections through the intrusion are located near its eastern end where contacts at both the roof and floor are exposed at several localities. Measured stratigraphic columns were constructed at three sites in this area, and their locations are shown in Fig. 1b. Specimens for thin sectioning were collected at vertical intervals of 2–5 m at each locality. Section A–A' contains an exposure of 135 m thickness in which the lower contact is visible but an unknown thickness of the upper part of the intrusion has been removed by erosion. Section B–B' contains a complete section of 150 m thickness through the intrusion, although the upper contact is exposed about 100 m south of the rest of the measured section. Both of these sections contain several texturally and petrographically distinctive sub-horizontal lithological layers that are described in detail below. The intrusion thins to the south and is only 24 m thick at section C–C', where it consists entirely of olivine diabase. The stratigraphic columns for sections A–A', B–B' and C–C' are shown in Fig. 2. Rock exposures between sections A–A' and B–B' are nearly continuous and the distinctive lithological units in this part of the intrusion can be correlated with confidence; however, the interval between these sections and section C–C' is not exposed and the pinching out of these units shown in Fig. 2 is speculative. The modal abundances of plagioclase, pyroxene, olivine and interstitial granophyre in specimens collected at section B–B' are shown vs stratigraphic height in Fig. 3. Modal analyses of representative specimens from the major lithological units within the intrusion are given with the bulk-rock chemical analyses in Table 1.
View larger version (56K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 2. Stratigraphic columns through the Graveyard Point intrusion showing the vertical distribution of rock types at measured sections A–A', B–B' and C–C' (Fig. 1). Correlations of rock units between sections A–A' and B–B' can be made with confidence, but units cannot be traced in the field between sections B–B' and C–C'. The vertical exaggeration is x7.
|
|
View larger version (36K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 3. Distribution of modal olivine, Ca-rich pyroxene, plagioclase, and interstitial granophyre in samples from section B–B'. Line A marks the first appearance of pyroxene as discrete subhedral crystals; line B indicates the base of the laminated gabbros.
|
|
Chilled borders of fine-grained, intergranular microgabbro are
well developed next to the upper and lower contacts wherever
they are exposed. Plagioclase (
An
81) and olivine (
Fo
79) are
the only phenocrysts in these rocks; they make up about 15–20%
of the mode and commonly occur together in glomeroporphyritic
clots up to 4 mm across (
Fig. 4a). In all sections, the lower
chilled border grades upward into fine- to medium-grained, dense,
black, subophitic- to ophitic-textured diabase in which olivine
and plagioclase are partly or completely enclosed by optically
continuous anhedral crystals of augite. Section C–C' is
composed entirely of this rock type, but in the thicker parts
of the intrusion (sections A–A' and B–B') a noticeable
change in the petrographic character of the rocks occurs at
about 10 m above the floor. At this height, augite appears for
the first time as discrete subhedral crystals and the rocks
become distinctly granular in texture, with augite and olivine
typically occurring in crystals up to 3 mm across and plagioclase
crystals generally being slightly smaller. This petrographic
boundary also marks a shift in the chemical compositions of
the rocks, which is discussed below. At
30 m above the lower
contact, the gabbros take on a moderately to strongly sub-horizontal
igneous lamination caused by the shape-preferred orientations
of augite and plagioclase (
Fig. 4b). These rocks are coarser
grained than the granular gabbros, with crystals typically being
5–8 mm in length. Augite and plagioclase are tabular in
shape and form compact frameworks of touching subhedral to euhedral
grains typical of cumulates (Irvine, 1982
). Imbricate textures
(tiling) are common, suggesting that the planar fabrics were
caused at least in part by laminar flow (e.g. Nicolas, 1992
;
Shelley, 1993
).
View larger version (132K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 4. Photomicrographs of representative specimens from the Graveyard Point intrusion; all were taken in plane-polarized light and are at the same scale, shown by the 5 mm bar in (a). (See text for detailed descriptions.) (a) Lower chilled border with small glomerocrysts of plagioclase and olivine; (b) laminated gabbro (arrow points to a granophyre vein that cuts across the plane of lamination); (c) ferrogabbro containing plagioclase, augite, skeletal Fe–Ti oxide, and pools of interstitial granophyre; (d) poikilitic gabbro consisting mainly of large, optically continuous augite enclosing small crystals of plagioclase.
|
|
Near the middle of the intrusion, about 55 m above the floor,
the rocks change from laminated gabbro to patchy textured coarse-grained
ferrogabbro having no preferred orientation of crystals. Some
zones within this unit contain crystals as much as 4 cm in length.
In thin sections the ferrogabbros are composed of a porous framework
of plagioclase, monoclinic pyroxene, olivine, apatite and magnetite
surrounding interstitial pools of fine-grained granophyre and
granophyre-lined vesicles (
Fig. 4c). The oxides commonly occur
as skeletal crystals. Within this unit are one or more lens-shaped
layers in which the granophyre forms as much as 50% of the total
rock. These exceptionally granophyre-rich zones, which have
the bulk composition of ferrodiorite, are easily recognized
in the field because they weather to a reddish brown color,
contain abundant vesicles, and have a distinctive brecciated
appearance (
Fig. 5a). The granophyre in these rocks is extremely
fine-grained and contains swallow-tailed or hollow-centered
crystals indicative of rapid crystallization (Lofgren, 1980
);
the coarse-grained parts consist of clusters of minerals that
are petrographically similar to those in the surrounding ferrogabbro.
View larger version (82K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 5. (a) Outcrop of mixed-textured rock composed of medium- to coarse-grained ferrogabbro containing veins and clots of fine grained siliceous granophyre (G). Scale bar represents 10 cm. (b) Area of intersecting sheets of siliceous granophyre cutting poikilitic gabbro in the upper part of the intrusion. Hammer is 36 cm long
|
|
Overlying the ferrogabbro and ferrodiorite is a sequence of
much less differentiated, poikilitic-textured olivine gabbros,
which typically consist of small (
1 mm) subhedral crystals of
plagioclase surrounded by much larger (
1 cm) anhedral poikilitic
crystals of augite or, less commonly, olivine (
Fig. 4d). This
rock type is not present in section A–A', where the upper
part of the intrusion has been removed by erosion. The poikilitic
textures of these rocks and their position in the upper part
of the intrusion suggest that they may form an upper border
series that crystallized
in situ near the roof. Where the upper
contact of the intrusion is preserved, the poikilitic-textured
gabbro is overlain by a few meters of dark gray, fine-grained,
olivine diabase similar to the rocks adjacent to the floor.
Distinct veins, sheets and pods of granophyre intrude all the other rock types except the olivine diabase in the lower 10 m of the intrusion. Granophyre is present as millimeter- to centimeter-wide veins in the laminated gabbro (Fig. 4b) and forms cross-cutting and conformable sheets as much as 30 cm thick in the upper one-third of the intrusion (Fig. 5b). At one locality, poikilitic gabbro is cut by an irregularly shaped, diapir-like, granophyre body about 5 m across. Medium- to coarse-grained granophyres are similar to the ferrogabbros in appearance and composition, and generally contain less than 52% SiO2. Fine-grained granophyres have higher silica contents, up to 68% in one dike, and contain small crystals of sodic plagioclase, ferroaugite, magnetite and apatite, ± pigeonite, ± fayalitic olivine, in a fine-grained matrix of intergrown quartz and alkali feldspar. Although discrete granophyre bodies are most abundant in the interior of the intrusion, a 1 m thick sheet of fine-grained granophyre also occurs for about 100 m along the roof of the intrusion, where it is in contact with hornfelsed country rock.
|
GEOCHEMISTRY
|
|---|
Analyses of bulk-rocks and small veins
About 100 samples from all parts of the intrusion and nine specimens
from the wall-rocks adjacent to the roof and floor were analyzed
for major elements and 14 trace elements by X-ray fluorescence
spectrometry (XRF) using the Rigaku 3370 system at the Washington
State University Geoanalytical Laboratory (the WSU laboratory
has since replaced this instrument). Sample preparation methods,
operating conditions and statistics for these analyses have
been described by (Johnson
et al., 1999
). Twenty-six samples
were analyzed for additional trace elements by inductively coupled
plasma mass spectrometry (ICP-MS) using the HP-4500+ instrument
at the WSU laboratory. Technical notes for the ICP-MS analyses
are available on the WSU Geoanalytical Laboratory web site.
Representative chemical analyses of bulk-rock specimens are
given in
Table 1, along with data on analytical precision. In
addition, the granophyre in several small (< 1 mm wide) veins
in thin sections of samples from the laminated gabbro and ferrogabbro
was analyzed for eight major oxides with the Cameca Camebax
electron microprobe at Washington State University. These analyses
were made using the Kevex spectrometer, with a beam width of
100–300 µm and an accelerating voltage of 20 kV.
Analyses of four samples from the chilled borders collected at various locations around the intrusion are given in Table 1 (analyses 1, 18, 19, 20). The compositions cluster at the boundary between olivine tholeiite and quartz tholeiite, according to the normative classification system of (Yoder & Tilley, 1962), and plot within the field of low-K, high-Ti transitional basalts (LKTB) on the MgO–TiO2–K2O diagram of (Hart et al., 1984). This composition is intermediate between mid-ocean ridge basalt (MORB)-like high-alumina olivine tholeiites (HAOT) from the northwestern margin of the Basin and Range province and the more evolved tholeiites typical of the younger (<3 Ma) basalts in the western Snake River Plain (SROT). The intrusion's chilled margin is, however, very similar in composition to the earliest erupted basalts (6–8 Ma) in the western Snake River Plain (White et al., 2002).
When all of the analyses of bulk-rock specimens and granophyre veins are plotted on an AFM diagram (Fig. 6a), they form a well-defined tholeiitic trend of iron enrichment followed by iron depletion and enrichment in alkalis. Concentrations of FeOT peak at around 19% in the ferrogabbros, well below the maximum iron enrichment of the Skaergaard liquid (McBirney, 1996), but similar to the maximum values found in volcanic glasses (e.g. Brooks et al., 1991) and close to the limit of iron enrichment predicted by the fractionation models of (Toplis & Carroll, 1996). Silica contents range between 45 and 51% for all analyzed rock specimens with the exception of those collected from the granophyre-enriched, mixed-textured ferrodiorites and the discrete granophyre intrusions (Table 1). Analyses recalculated to CMAS end-members are plotted on the 1 atmosphere Ol–Di–Q pseudoternary diagram in Fig. 6b. The chilled borders and ophitic-textured olivine diabases plot within the field of olivine (+plagioclase), the granular-textured augite-bearing gabbros in the interior of the intrusion (including the laminated rocks) plot along the olivine–Ca-rich pyroxene boundary or within the field of Ca-rich pyroxene (+plagioclase), and the granophyres plot along the pigeonite–Ca-rich pyroxene boundary (+plagioclase). Samples of diabase from the lower part of the intrusion plot closer to the Ol corner than the chilled border samples, which is consistent with the small spike in modal olivine at about 5 m above the floor of the intrusion (Fig. 3) and may indicate that these rocks contain excess olivine. The laminated gabbros plot closer to the Di corner than any of the other specimens, which is consistent with their high content of modal augite (Fig. 3) and suggests that they may contain excess pyroxene.
Chemical profiles for MgO, Ni, Sc and Zr are shown for measured
section B–B' in
Fig. 7. A notable feature in these profiles
is the chemical discontinuity at about 10 m above the base of
the intrusion, coincident with the boundary between the ophitic-textured
olivine diabase and the coarser-grained, granular-textured olivine
gabbro. Rocks immediately above this boundary contain less MgO,
substantially less Ni, and somewhat more Sc and Zr than the
rocks below it. The behavior of other oxides and trace elements
across this discontinuity can be seen by comparing analyses
3 and 4 in
Table 1. XRF analyses of seven specimens from the
lower part of section A–A' show a similar discontinuity
about 13 m above the lower contact. Concentrations of MgO and
Ni abruptly increase again near the top of the section at the
contact between the poikilitic gabbros and the olivine diabase
adjacent to the roof. It is clear from these profiles that the
average composition of the intrusion at this section does not
correspond to the compositions of the chilled borders. The lack
of mass balance within this part of the intrusion combined with
the abruptness of the lower and upper chemical discontinuities
can be explained by the addition of more evolved magma into
a sheet initially formed by a pulse of more mafic magma. The
lack of internal chilled borders indicates that the second influx
of magma must have occurred while the initial sheet was still
hot or partly molten.
The strongly incompatible trace elements Zr, Ba, Rb, Nb, La,
Th, Hf and Ta maintain nearly constant ratios with one another
throughout the suite of rock types within the intrusion. Plots
utilizing these elements result in straight-line trends that
project to the origin (
Fig. 8), while the volcanic and sedimentary
rocks adjacent to the intrusion generally plot well away from
these trends. Although small degrees of contamination by the
shallow crust cannot be ruled out, these relations are consistent
with the granophyres and other evolved rocks within the intrusion
having formed by fractional crystallization of a magma or magmas
similar in composition to the chilled border. None of the granophyres,
including the 1 m thick sheet adjacent to the roof, appear to
be partial melts of the surrounding country rock.
View larger version (14K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 8. Variation of Rb vs Zr (ppm) for all analyzed bulk-rock samples from the intrusion compared with analyzed samples of silicic tuffs and sediments from the surrounding country rocks. Symbols for samples from the intrusion are the same as in Fig. 6; country rocks are indicated by open stars.
|
|
Mineral compositions
The compositions of olivine, pyroxene and plagioclase in selected
specimens from the intrusion were determined using a Cameca
Camebax electron microprobe at Washington State University using
a 20 kV current and a 4 µm beam width. Additional microprobe
analyses of plagioclase were obtained at the University of Oregon
using a Cameca SX50. Typically, the compositions of about 8–10
crystals each of plagioclase, pyroxene and olivine (if present)
were determined for a specimen, and between one and three points
were analyzed in the interior of each crystal. Representative
mineral analyses are included in Electronic Appendix 1 at
http://petrology.oxfordjournals.org.
Variations in the compositions of the major silicate phases with height through section B–B' are shown in Fig. 9. A notable feature in this diagram is the abrupt shift in the compositions of olivine and plagioclase that takes place at the same stratigraphic height as the bulk-rock chemical discontinuity (line A in Fig. 9). Pyroxene compositions are not displaced at this boundary; however, this mineral occurs only as interstitial grains or poikilitic plates in the lower 10 m of the section and was probably not a primocryst. The compositions of plagioclase and pyroxene in the laminated gabbros are noteworthy for their relatively constant average values throughout the unit and for the small compositional variations among crystals in a single thin section. The uniform compositions of the minerals in these rocks supports the interpretation that variations in the bulk-rock chemistry of samples from this unit are probably due to differences in the amounts and proportions of excess crystals.
View larger version (32K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 9. Variations in the compositions of olivine, pyroxene and plagioclase in rocks sampled from measured section B–B'. Most of the points are averages of microprobe analyses of 8–10 crystals in a single thin section; error bars are based on 1 SD. Lines A and B are drawn at the textural and chemical discontinuities described for Figs. 3 and 7.
|
|
Isotopic ratios
Sr isotopic ratios were determined for 12 samples from the Graveyard
Point intrusion and four specimens of wall-rock by thermal ionization
mass spectrometry at Miami University (Ohio) (
Table 2). Measured
analyses were corrected to
87Sr/
86Sr = 0·70800 for the
E&A SrCO
3 standard, and initial ratios were calculated for
an age of 6·7 Ma. The 12 analyzed rocks from the intrusion
yielded initial
87Sr/
86Sr ratios between 0·7059 and 0·7065,
with specimens of the Ni-rich olivine diabase having slightly
lower values than those from the interior of the intrusion (
Fig. 10).
On the other hand, there are no systematic differences between
the Sr isotopic ratios of the granophyre sheets and those of
the more mafic gabbro and ferrogabbro. The silicic pyroclastic
rocks and tuffaceous sediments that surround the intrusion display
a greater range of Sr ratios (0·7048–0·7067),
which increase with increasing silica content. Taken together,
the Sr isotopic analyses support the interpretation that the
granophyric liquids formed by differentiation of the more mafic
magmas after the final emplacement of the intrusion and without
any significant contribution from the wall-rocks.
|
DISCUSSION
|
|---|
Emplacement of the intrusion
Several types of field evidence indicate that the Graveyard
Point intrusion was emplaced at a shallow depth. These include
the presence of well-developed, fine-grained chilled borders,
the abundance of round vesicles in some parts of the upper chilled
border, and the miarolitic texture of evolved rocks in the interior
of the intrusion. Moreover, XRD studies have shown that tridymite
is present in fused siltstone within a few centimeters of the
contacts (Ferns, 1989
). This occurrence allows the maximum pressure
at the site of emplacement to be estimated using the pressure–temperature
values for the thermodynamically and experimentally determined
ß-quartz–tridymite transition (Tuttle &
Bowen, 1958
; Berman, 1988
). An estimate of the maximum temperature
of the first magma injected into the intrusion can be made using
the MELTS (Ghiorso & Sack, 1995
) and MIXFRAC (Nielsen, 1988
)
computer fractionation models. These programs yield temperatures
of 1185°C and 1205°C, respectively, for the first appearance
of olivine + plagioclase in a liquid with the composition of
the chilled border. A temperature of 1096°C is obtained
for the chilled border magma using the plagioclase + liquid
thermometer of (Putirka, 2005
), although this value is substantially
lower than temperatures estimated for SRP basalts in previous
studies (Leeman & Vitaliano, 1976
; Honjo & Leeman, 1987
).
If a temperature of 1180°C is assumed for the magma emplaced
by the first pulse, then a maximum pressure of 175 MPa can be
calculated from the relation
P (MPa) = 0·561
T (°C)
– 487 ± 6 for the quartz–tridymite transition
(Hirschmann
et al., 1997
). This value corresponds to a maximum
depth of around 6·5 km, given an average density of 2700
kg/m
3 for the overlying rock. (Putirka, 2005
) also formulated
an expression for calculating pressure based on plagioclase
+ liquid equilibria, which, when applied to the mineral and
bulk-rock analyses for the chilled border, yields a value of
129 MPa (equivalent to
4·8 km). Because this would be
the pressure at which plagioclase equilibrated with the liquid
prior to injection, a shallower depth can be inferred for the
intrusion.
The simplest explanation for the lack of mass balance between the composition of the chilled border and the bulk composition of the intrusion is that it was formed by more than one injection of magma. The abrupt shift in rock and mineral compositions at about 10 m above the base of the intrusion is inferred here to mark the lower contact of the second magma pulse. The initial influx of tholeiitic magma contained small phenocrysts of olivine and plagioclase, as indicated by thin sections of the chilled border (Fig. 4a). The second pulse was more voluminous and more chemically evolved, and carried with it a substantial proportion of medium- to coarse-grained crystals of augite and plagioclase, and a smaller amount of olivine. The depth at which this magma originated can only be roughly estimated because its composition is not known with any certainty; however, experiments utilizing basalts from the Snake River Plain indicate that Ca-rich pyroxene is not likely to be a near-liquidus phase at pressures less than 500 MPa (Thompson, 1972; Leeman & Vitaliano, 1976). The well-defined layer of laminated gabbro beginning about 20 m above the presumed base of the second magma pulse (30 m above the floor of the intrusion) probably represents a crystal-rich tongue formed by flow differentiation during the ascent and lateral injection of the magma (e.g. Simkin, 1967; Upton & Wadsworth, 1967; Komar, 1972; Richardson, 1979; Husch, 1990; Mangan et al., 1993). The imbricate textures in these rocks and the wedge-shaped geometry of the unit support this interpretation. Because the laminated gabbro contains variable proportions of excess augite and plagioclase, chemical profiles through it do not show the well-developed D-shaped enrichments for MgO that have been noted in sills containing similarly formed concentrations of olivine and orthopyroxene (e.g. Gibb & Henderson, 1992; Marsh, 1996).
It is not entirely clear when the magma that formed the poikilitic gabbro in the upper part of the intrusion in section B–B' was emplaced. The Sr isotopic ratios of these rocks are more like those of samples from the interior of the intrusion than they are to analyses of the olivine diabase (Fig. 10); however, evidence discussed below indicates that the poikilitic gabbros are more probably related to the initial influx of magma. Although these rocks are less mafic than the chilled border and the basal olivine diabase, they contain more MgO and Ni and have lower contents of excluded elements than any of the other rocks in the intrusion, including the laminated cumulates (Table 1). Moreover, the small primocrysts of plagioclase in the poikilitic gabbros are more calcic than plagioclase in the laminated rocks (Fig. 9). These relations are the reverse of what is generally observed in intrusions where an upper border series and a lower cumulate series are coeval (e.g. Naslund, 1984). It is more likely that small crystals of plagioclase and olivine carried in with the first pulse of magma settled out of the upper part of the intrusion and accumulated in the lower part, resulting in formation of the olivine diabase. Major-element based least-squares calculations (Bryan et al., 1969) support this hypothesis: analyses of the poikilitic gabbro can be closely approximated by subtracting 8–12% olivine + plagioclase in roughly equal amounts from the chilled border composition, and the olivine diabase can be duplicated by adding 4–11% of these minerals to the analysis of the chilled border. The sums of the squares of the residuals range from 0·02 to 0·21 when the compositions of olivine and plagioclase in the chilled border are used in the calculations. The redistribution of crystals in the amounts and proportions indicated by the least-squares calculations is plausible given that the average phenocryst content in samples of the chilled border is around 5% olivine and 13% plagioclase. If this interpretation is correct, then almost all of the olivine phenocrysts and about half of the plagioclase crystals would have been removed from the upper–middle part of the intrusion and concentrated in the lower 10 m. However, given that the zone of poikilitic gabbro in section B–B' is substantially thicker than the layer of olivine diabase near the floor, it is surprising that the diabase samples do not contain an even greater proportion of accumulated crystals. This discrepancy may be due in part to insufficient sampling of the lower diabase. Alternatively, residual liquid left over after the separation of the olivine and plagioclase phenocrysts may have migrated along a sloping roof and become concentrated in a cupola, in which case the poikilitic gabbro would not be as voluminous elsewhere in the intrusion as it appears to be from section B–B'.
Considering the discussion above, the following sequence of events is proposed for the emplacement of the Graveyard Point intrusion.
- Basaltic magma containing about 18% small phenocrysts of olivine and plagioclase was injected into a sequence of silicic volcanic rocks and tuffaceous sediments, forming an irregularly shaped but generally sheet-like intrusion that was as much as 60 m thick in the area of measured section B–B' but thinned to about 25 m at section C–C'.
- Following the formation of the chilled borders, the phenocrysts of olivine and, to a lesser extent, plagioclase that were carried in with the magma settled and accumulated in the lower 10 m of the intrusion, producing a layer of MgO- and Ni-enriched olivine diabase. Crystal-depleted liquids in the upper part of the intrusion may have become concentrated in one or more cupolas in the roof where they solidified in situ, forming the distinctive poikilitic textured gabbros.
- A second injection of more evolved basaltic magma was emplaced into the thickest part of the intrusion after the initial pulse had mostly, but probably not completely, crystallized. It did not invade the thinner distal part of the original intrusion (section C–C'), perhaps because those areas were already solid. The second pulse carried in a large proportion of augite and plagioclase phenocrysts and inflated the intrusion to at least 150 m thick in the area of sections A–A' and B–B'. The phenocrysts were concentrated by flow differentiation during the ascent and emplacement of the magma, and formed a wedge-shaped layer of crystal mush between 60 and 25 m thick.
Differentiation processes
The chemical composition of magma in the leading edge of the first pulse is preserved in the chilled border, and, as discussed above, the primary process causing this magma to differentiate after emplacement appears to have been the redistribution of olivine and plagioclase phenocrysts. The composition of the second magma is not as well constrained because it did not form distinctive chilled borders; however, the granular-textured gabbro just above the chemical discontinuity in section B–B' may provide a reasonable approximation (sample GP-338, Table 1, column 4). If this is assumed to be true, then the magma injected in the second pulse was a quartz tholeiite with SiO2 between 49 and 50% and MgO between 5 and 6%. The lack of internal chilled borders implies a short time interval between the emplacement of the two magmas; for this reason, a simple scenario in which the second magma evolved from the first during the time between injections is probably unrealistic. It is more likely that they were derived from different parts of a complex system of interconnected reservoirs that had undergone differing amounts of fractional crystallization of the same or a similar parent magma; in other words, from different regions and probably from different depths within the magmatic mush column.
There is no evidence to indicate that any magmas with compositions more evolved than quartz tholeiite were added from outside the intrusion, and the geochemical and isotopic data preclude significant amounts of wall-rock contamination. Therefore, the ferrogabbros, mixed-textured ferrodiorites and silicic granophyres all must have evolved within the intrusion itself. Using the analysis of sample GP-338 as a proxy for the bulk composition of the second magma, enrichments of the most strongly excluded trace elements (Zr and Th) indicate that 20–40% fractional crystallization would be needed to generate the range of compositions in the ferrogabbros, and about 65–75% crystallization is necessary to form the silicic granophyres. The behavior of the incompatible minor and trace elements in these rocks is shown on the normalized trace-element diagram in Fig. 11, along with a specimen of the laminated gabbro. Element concentrations were normalized to the analysis of sample GP-338. The typical ferrogabbro is enriched in nearly all of the 22 elements on the diagram relative to its inferred parent magma, with the notable exception of Sr, which is enriched in the laminated gabbro. Although not shown in Fig. 11, Sc and Cr are also depleted in the ferrogabbros and enriched in the laminated rocks relative to GP-338 (Table 1). The trace-element relations are consistent with fractionation of plagioclase + augite ± olivine, the minerals that appear to have been concentrated in the laminated gabbro. The mixed-textured ferrodiorite and the silicic granophyre are progressively more enriched in most of the incompatible elements but also produce increasingly strong negative anomalies for Sr and Ti (Fig. 11). Although the ferrodiorite is enriched in P relative to the ferrogabbro, the granophyre is strongly depleted in this element, even compared with the inferred composition of the parent magma. The variations in Sr, Ti and P in these rocks are consistent with a fractionation scheme that initially included plagioclase and progressively added Ti–Fe oxides and apatite. This does not mean to say that these minerals settled out of a compositionally evolving liquid; it is more likely that liquids separated from frameworks of crystals in which these minerals were progressively crystallizing.
View larger version (39K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 11. Incompatible element diagram for representative specimens of laminated gabbro, ferrogabbro, ferrodiorite and granophyre. Analyses are normalized to GP-338 (Table 1), which is inferred to approximate the bulk composition of the second magma emplaced into the intrusion.
|
|
The physical processes that result in the segregation of chemically
evolved liquids have been the focus of many of the more recent
studies of small to moderate-sized mafic intrusions and thick
lava flows. (Philpotts
et al., 1996
) and (Marsh, 2002
) emphasized
a number of important differences between the processes that
form low-Si, generally coarse-grained gabbroic segregations
(e.g. ‘gabbroic pegmatites’ of Wager & Deer,
1938
; ‘dolerite pegmatites’ of Walker, 1949
) and
those that form Si-rich, commonly fine-grained granophyres (e.g.
Hotz, 1953
). Following their lead, these rock types are discussed
separately below, beginning with the ferrogabbros. Where they
are present in the Graveyard Point intrusion, the ferrogabbros
are similar in texture and composition to the coarse-grained,
iron-rich (but not silica-rich) gabbroic pegmatites or pegmatitic
segregation veins commonly found in diabase sills, solidified
lava lakes and thick basalt flows. With few exceptions, studies
of these bodies have attributed their origin to the upward migration
and segregation of liquids formed within the lower crystal mush
after 20–30% crystallization (e.g. Puffer & Horter,
1993
; Philpotts
et al., 1996
; Mitchell
et al., 1997
). Because
these liquids are commonly iron-rich and therefore dense, various
mechanisms have been suggested that would enhance their buoyancy
and facilitate their movement upward through the crystal pile.
Shirley (1986
, 1987
) emphasized the role of crystal compaction
to redistribute interstitial liquids in the Palisades Sill,
and compaction-driven upward movement of evolved low-silica
melts has been documented in thick basalt flows (Philpotts
et al., 1996
; Philpotts & Philpotts, 2005
). (Larsen & Brooks,
1994
) suggested that gabbroic pegmatites in the Skaergaard intrusion
formed by the convective rise of interstitial liquids whose
densities had been reduced by dissolved water. (Puffer &
Horter, 1993
) attributed the pegmatitic segregation veins in
flood basalts to the transport of interstitial liquids in segregation
vesicles, a process described by (Helz
et al., 1989
) for the
transfer of evolved melts in the Kilauea Iki lava lake. This
model is based in part on the process described as ‘gas
filter pressing’ by (Anderson
et al., 1984
) and as ‘vapor-differentiation’
by (Goff, 1996
), by which interstitial melts are forced into
vesicles formed when magmatic gases exsolve in response to crystallization.
The evolved liquid is transported upward in the buoyant mixture
of vapor- and melt-filled bubbles.
Although gas vesicles are common in the evolved mafic rocks of the Graveyard Point intrusion, they are not present in the underlying laminated gabbros. This suggests that exsolution of the gas phase took place after the ferrogabbroic liquids became segregated. Moreover, most of the ferrogabbros are depleted in Ba, Rb and Zr relative to values predicted by major-oxide based mass-balance calculations, whereas these elements should be preferentially enriched in melts that were transported upward in volatile-rich segregation vesicles (Puffer & Horter, 1993). If a gas transport mechanism did not play an important role at this stage, then it is likely that the evolved mafic liquids migrated upward in response to compaction of the crystals within the laminated unit. Although iron-rich, the buoyancy of these liquids would have been enhanced by increasing amounts of dissolved magmatic gases as crystallization proceeded within the crystal mush. In addition, field relations indicate that slabs of partly or wholly crystallized rock became detached from the lower part of the poikilitic gabbro that had crystallized next to the roof of the intrusion, causing the stratigraphic complexity observed in section B–B' (Fig. 2). Downward movement of this slab of largely crystalline crust would have augmented compaction in the crystal mush and caused the interstitial liquids to be drawn into the tear in a manner described by Marsh et al. (1991) and (Carman, 1994).
The silicic granophyres are mainly concentrated in the upper third of the intrusion, where they occur as dikes, sheets and pods within the poikilitic gabbro. As noted above, the low TiO2 and P2O5 contents of the silicic granophyres complement the high values of these oxides in many of the ferrogabbros. This relationship can be explained by the movement of late-stage, interstitial liquids out of a framework of crystals in which apatite and Fe–Ti oxides were retained. Segregation of these liquids was probably driven by the vapor-differentiation process described above. The abundance of irregularly shaped miarolitic cavities in the mixed-textured ferrodiorite and the presence of granophyre-lined vesicles in the ferrogabbro (Fig. 12) are compelling evidence in support of this process. The brecciated appearance of the mixed-textured ferrodiorites (Fig. 5a) suggests that vesiculation may have been very rapid and was perhaps initiated by decompression triggered by fracturing of the overlying country rocks. An abrupt drop in vapor pressure may also explain the quench textures of crystals in the interstitial granophyres in many of the ferrodiorites.
View larger version (176K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 12. Photomicrograph (plane-polarized light) of a granophyre-lined vesicle in ferrogabbro from the interior of the Graveyard Point intrusion. The center of the vesicle is filled with limonite (lim). Small crystals of feldspar in the surrounding granophyre display quench textures and have compositions around Ab75Or10An15. Coarse-grained crystals are plagioclase (pl), augite (aug) and opaque oxides. Scale bar represents 1 mm.
|
|
A two-stage process is therefore suggested in which iron-rich
mafic liquids were formed by localized fractionation within
the tongue of crystals emplaced with the second magma pulse
and were expelled by crystal compaction. Further
in situ crystallization
of this liquid produced a second generation of more evolved,
silica-enriched, liquids, which in turn separated from their
own rigid, but porous, crystal framework in response to the
exsolution and migration of bubbles of magmatic vapor. The combined
process is generally similar to that described by (Philpotts
et al., 1996
) for the differentiation of the thick Holyoke lava
flow, although compelling evidence does not exist at the Graveyard
Point intrusion to support their suggestion that the late-stage
silica-rich granophyres form from immiscible liquids.
Liquid trend
Although cumulus processes in small intrusions are generally inefficient compared with those in larger bodies, some of the rocks in the Graveyard Point intrusion clearly contain excess amounts pyroxene and/or plagioclase, and therefore analyses of these specimens cannot be considered to represent any liquid. On the other hand, analyses of the small sheets and dikes of granophyre must represent late-stage liquid compositions, and many of the ferrodiorites and ferrogabbros are probably close to the compositions of liquids that existed at different places and times within the intrusion. Analyses of these rocks are plotted on Fenner diagrams in Fig. 13 to illustrate the trend of liquids inferred to have evolved within the intrusion after the second magma was emplaced. The laminated and granular gabbros from the lower middle part of the intrusion have been excluded from these plots, and they also do not include any data for the mafic rocks related to the first influx of magma. The chemical trend for the late-stage liquids is compared with model liquid trends produced by the fractional crystallization simulations of the experimentally constrained MIXFRAC program of (Nielsen, 1988). The analysis of the first sample above the chemical discontinuity in section B–B' (GP-338, Table 1, column 4) was used as the composition of the parent magma in the computer model and is indicated by a star in Fig. 13. The calculations were made with the oxygen fugacity fixed at the fayalite–magnetite–quartz (FMQ) buffer. The low-pressure trend produced by MIXFRAC broadly mimics the sample compositions and inflection points on the Fenner diagrams, but does not reach the level of iron enrichment observed in the rocks. The poor fit for FeOt may be due to the inability of computer fractionation models to accurately predict the saturation temperature of Fe–Ti oxides in iron-rich magmas [see the discussion by (Toplis & Carroll, 1996)]. In general, however, the model is consistent with the liquid trend inferred from the bulk-rock analyses, and supports the interpretation that, following the emplacement of the second pulse of magma, the remaining liquid evolved within the intrusion by some process of crystallization-driven differentiation.
View larger version (24K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 13. Variation of major and minor oxides (wt%) plotted against MgO (wt%) for ferrogabbros, ferrodiorites and granophyres from the Graveyard Point intrusion. [See text for discussion of the low-pressure model liquid trend derived from MIXFRAC (Nielsen, 1988).]
|
|
Comparison with Snake River Plain lavas
The late Miocene to Pleistocene basalts of the western SRP erupted
from as many as 400 different vents associated with scoria cones,
small shields and phreatomagmatic centers. All but the very
youngest of these lavas plot in the field of tholeiitic rocks
on AMF diagrams (
Fig. 14) and display trends of increasing FeO
T and TiO
2, and slightly decreasing SiO
2, as MgO decreases. To
date, only one tholeiitic lava flow has been identified in the
western SRP that appears to have differentiated to the point
where SiO
2 enrichment had begun (Bonnichsen & Godchaux,
2002
). Maximum concentrations for FeO
T and TiO
2 in the western
SRP volcanic suite are about 17·5% and 4·5%, respectively.
Similar values are recorded for many of the ferrogabbros in
the Graveyard Point intrusion, although a few specimens contain
as much as 20% FeO
T and 5·5% TiO
2, possibly owing to
small amounts of excess magnetite.
Ratios of CaO/Al
2O
3 in both suites initially increase slightly
with differentiation and then decline (
Fig. 15). Although this
trend is not particularly strong in the volcanic suite, it does
suggest that small amounts of Ca-rich pyroxene were removed
from these magmas. This interpretation is in apparent conflict
with the observation that pyroxene-phyric lavas have never been
reported from the western SRP, although the Graveyard Point
intrusion provides compelling evidence that pyroxene can be
an abundant crystallizing phase at some depth within the magmatic
mush column.
View larger version (23K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 15. Selected oxide and trace-element ratios plotted against Mg-number (Mg-) for analyzed rocks from the Graveyard Point intrusion (open circles) and lavas from the western SRP (crosses). The average K2O/P2O5 ratio in rocks from Layered Series Upper Zone b of the Skaergaard intrusion (Skg UZb; McBirney, 1996) is shown for comparison. (See text for discussion.)
|
|
The greatest differences between the compositions of the SRP
lavas and those of the intrusion occur in the behaviors of the
incompatible minor oxides and trace elements. Although abundances
of these elements are about the same in the least differentiated
rocks of each suite, at moderate degrees of differentiation
(Mg numbers
40–50) their concentrations are much greater
in the lavas than they are in the plutonic rocks. The enrichment
of P
2O
5 in the volcanic suite is particularly noteworthy, as
it exceeds that of either K
2O or TiO
2. Among the incompatible
trace elements, Y and the heavy REE are only slightly enriched
in the ferrobasalts compared with the strong increases in the
concentrations of Zr, Sr, Ba, La and Nb (
Fig. 15). The behaviors
of the excluded elements in the lavas are best explained by
a process of fractional crystallization combined with open-system
mixing with a source rich in excluded elements and particularly
enriched in phosphorus. This source was probably not within
the granitic rocks of the Idaho Batholith because most of them
have high K/P ratios. A similar argument can be used to rule
out the Miocene age SRP rhyolites erupted just before the beginning
of western SRP basalt magmatism. An older cratonic source is
unlikely because Sr-isotopic analyses for western SRP lavas
indicate that
87Sr/
86Sr ratios are lower for the high P
2O
5 ferrobasalts
than they are for less evolved basalts in the same region (White
et al., 2002
).
A similar trend of P2O5 enrichment and decreasing K2O/P2O5 was noted by (Geist et al., 2002) for basaltic lavas sampled in drill core at the Idaho National Laboratory in the eastern Snake River Plain. They proposed that these magmas evolved by combined fractional crystallization and assimilation (AFC) of high-P2O5 ferrogabbro contained within a differentiated mafic intrusion emplaced at mid-crustal levels below the eastern SRP. A similar model is proposed here to explain why the trace element trends in the western SRP lavas are so different from those in the Graveyard Point intrusion. The high incompatible element abundances and variable but generally low K2O/P2O5 ratios in the lavas are consistent with mixing of a crystallizing basaltic magma with partial melts of ferrogabbro and granophyre in layers or veins within a differentiated mafic intrusion. Partial melts of ferrogabbros containing a few percent cumulus apatite would be enriched in phosphorus; however, if some apatite were retained in the melting residuum, the bulk distribution coefficients for Y and Yb would be greater than those for elements such as Zr and La, which are excluded from all likely minerals in the residuum including apatite. There is good evidence suggesting that mafic intrusions may be common beneath the western SRP: fast seismic velocities at mid-crustal levels beneath the western plain and high gravity anomalies along the axis of the plain have both been attributed to the presence of mafic intrusions in the mid-crust (Mabey, 1982; Wood & Clemens, 2002). The Graveyard Point intrusion demonstrates that even relatively small western SRP sills can contain ferrogabbros with relatively high abundances of P2O5 (>1·5 wt%) and sheets of granophyre enriched in excluded trace elements (Zr >700 ppm; Ba >1500 ppm). More extreme enrichment of P2O5 might be expected in larger intrusions, where differentiation is likely to be more efficient (for example, Upper Zone b in the Layered Series of the Skaergaard intrusion).
|
SUMMARY AND CONCLUSIONS
|
|---|
Variations in the textures and compositions of rocks and minerals
composing the Graveyard Point intrusion can be explained by
multiple intrusions of magma followed by shallow-level
in situ crystallization, compaction and liquid migration. The initial
injection of magma consisted of olivine tholeiite, which is
preserved in chilled borders next to both the roof and floor
of the intrusion. Redistribution of existing phenocrysts of
olivine and plagioclase produced a lower zone of olivine diabase
enriched in MgO relative to the chilled border, and an upper
zone of poikilitic-textured gabbro that is depleted in MgO.
Injection of a second magma caused some parts of the intrusion
to inflate to as much as 150 m in thickness. In contrast to
the pattern of re-injection observed in many differentiated
sills, the second pulse was more evolved than the resident magma,
although it also entrained large amounts of plagioclase and
augite phenocrysts. Rock textures and chemical profiles suggest
that these phenocrysts were concentrated by flow differentiation
into a wedge-shaped horizon of distinctive laminated gabbro
as much as 60 m thick. Following emplacement of the second magma,
interstitial ferrogabbroic liquids separated from the crystal
mush, which compacted, and accumulated beneath the earlier formed
poikilitic-textured gabbro of the roof zone. Crystallization
of the ferrogabbroic magma in turn produced granophyric liquids,
which migrated out of interstitial spaces, probably in response
to oversaturation and exsolution of dissolved volatiles. Upward
movement of this low-density mixture of evolved granophyric
liquid and exsolved vapor inflated and fractured the uppermost
zones of the partially crystallized ferrogabbro, and sheets
of granophyric liquid intruded upward into the poikilitic gabbro.
The complex origin inferred for the Graveyard Point intrusion supports the recent observation by (Gibb & Henderson, 2006) that ‘with few exceptions’, detailed investigations of large mafic sills indicate a history of multiple intrusions. The substantial differences in the chemical compositions and phenocryst contents of the two pulses of magma (but not in their excluded trace element ratios) suggest that they evolved in different parts, and probably at different depths, within an interconnected magmatic mush column similar to that described by (Marsh, 2004). The extraordinary degree of differentiation observed in this relatively small intrusion is attributed to a combination of factors including the emplacement of a large proportion of crystals with the second magma pulse and a relatively high volatile content in the second-stage liquid.
The Graveyard Point intrusion provides the only documented example of the extended differentiation of a western SRP magma in a shallow pluton. Comparison of the major element chemical trends inferred for the Graveyard Point liquid with those produced by western SRP lavas shows that they are generally similar. This supports the interpretations of many researchers that SRP tholeiites have evolved by fractional crystallization at relatively low pressures (e.g. Leeman & Vitaliano, 1976). In contrast, the trends for P2O5 and some trace element ratios such as La/Yb are substantially different for the lavas compared with the intrusive suite. These differences are attributed to the interaction of rising SRP magmas with apatite-bearing mafic cumulates and evolved granophyric liquids at various levels within the crust.
|
SUPPLEMENTARY DATA
|
|---|
Supplementary data for this paper may be obtained at
Journal of Petrology online.
|
ACKNOWLEDGEMENTS
|
|---|
I am grateful to Gregg Beukelman and Doug Brown for their help
in the field, and to Bill Hart for generously providing the
Sr isotopic analyses. Discussions on the outcrop with Bill Bonnichsen,
Bill Hart and Gregor Markl were very helpful. I thank Bruce
Marsh, Bill Leeman and Wendy Bohrson for their insightful reviews
and helpful suggestions, which greatly improved this paper.
This research was supported by a grant from the National Science
Foundation (EAR-9219127).
*Corresponding author.Telephone: 208-426-3633. E-mail: cwhite{at}boisestate.edu
TOP
ABSTRACT
INTRODUCTION
