petrogenesis and implications of calc- alkaline cryptic hybrid … · washburn volcano, absaroka...
TRANSCRIPT
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 PAGES 663–703 2002
Petrogenesis and Implications of Calc-Alkaline Cryptic Hybrid Magmas fromWashburn Volcano, Absaroka VolcanicProvince, USA
T. C. FEELEY1∗, M. A. COSCA2 AND C. R. LINDSAY1
1DEPARTMENT OF EARTH SCIENCES, MONTANA STATE UNIVERSITY, BOZEMAN, MT 59717, USA2INSTITUTE OF MINERALOGY AND GEOCHEMISTRY, UNIVERSITY OF LAUSANNE, BFSH 2, 1015 LAUSANNE,
SWITZERLAND
RECEIVED JUNE 16, 2000; REVISED TYPESCRIPT ACCEPTED OCTOBER 18, 2001
KEY WORDS: petrogenesis; magma mixing; calc-alkaline; Absaroka Vol-The petrogenesis of calc-alkaline magmatism in the Eocene Absarokacanic Province; 40Ar/39Ar datesVolcanic Province (AVP) is investigated at Washburn volcano, a
major eruptive center in the low-K western belt of the AVP. New40Ar/39Ar age determinations indicate that magmatism at the volcanocommenced as early as 55 Ma and continued until at least 52 Ma.Although mineral and whole-rock compositional data reflect nearequilibrium crystallization of modal phenocrysts, petrogenetic mod-
INTRODUCTIONeling demonstrates that intermediate composition magmas are hybridsformed by mixing variably fractionated and contaminated mantle- The widespread, but poorly understood Challis–derived melts and heterogeneous silicic crustal melts. Nd and Sr Absaroka volcanic episode affected large areas of theisotopic compositions along with trace element data indicate that northwestern USA in the Eocene following Laramidesilicic melts in the Washburn system are derived from deep-crustal crustal shortening (i.e. the ‘Challis arc’; Armstrong, 1978).rocks broadly similar in composition to granulite-facies xenoliths in Volcanic rocks associated with this event are particularlythe Wyoming Province. Our preferred explanation for these features voluminous in the Absaroka Volcanic Province (AVP)is that mantle-derived basaltic magma intruded repeatedly in the of northwestern Wyoming and southwestern Montana,deep continental crust leading to fractional crystallization, silicic USA, where >20 000 km3 of calc-alkaline to shoshoniticmelt production, and homogenization of magmas, followed by ascent magmatic rocks are exposed in the Absaroka, Gallatin,to shallow reservoirs and crystallization of new plagioclase-rich and Beartooth Ranges (Fig. 1; Absaroka Volcanic Super-mineral assemblages in equilibrium with the intermediate hybrid group of Smedes & Prostka, 1972). Because the Absarokaliquids. The implications of this process are that (1) some calc- volcanic rocks have long been presumed to have calc-alkaline magmas may only be recognized as hybrids on purely alkaline compositional affinities and across-strike en-chemical grounds, particularly in systems where mixing precedes richments in K2O similar to magmas erupted in someand is widely separated from crystallization in space and time, and continental volcanic arcs (e.g. Dickinson & Hatherton,(2) given the role ascribed to crustal processes at Washburn volcano, 1967; Chadwick, 1970), early workers attributed theirthe variation between rocks that follow calc-alkaline trends in the origin to shallow subduction of the Farallon plate beneathwestern AVP and those that follow shoshonitic trends in the east the North American plate during the Eocene (e.g. Lipmancannot simply reflect higher pressures of fractionation to the east in et al., 1972). However, this interpretation is now con-
troversial for several reasons. First, spatial and temporalMoho-level magma chambers in the absence of crustal interaction.
Extended dataset can be found at http://www.petrology.oupjournals.org†∗Corresponding author. Telephone: 406/994-6917. Fax: 406/994-6923. E-mail: [email protected] Oxford University Press 2002
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
studies have been unable to decipher any logical time- TECTONIC AND GEOLOGICtransgression of magmatic activity throughout the north-
SETTINGwestern USA during the Eocene, with the result that theExisting geochronologic information indicates that the‘Challis arc’ was much wider than any modern arc andbulk of the magmas in the AVP were erupted betweenoriented perpendicular to the present plate margin (Fig.54 and 38 Ma, following or possibly overlapping the1). On the basis of this distribution and geochemicallatest phases of Laramide foreland thrusting, which endedarguments some workers now consider the Challis–at>55 Ma at this latitude (Love et al., 1975; Armstrong,Absaroka volcanic episode as resulting mainly from1978; Feeley et al., 1999; Hiza, 1999). Eruption of theregional lithospheric extension and resultant de-Absaroka rocks also appears to temporally coincide withcompression melting (Dudas, 1991; Hooper et al., 1995;the onset of regional crustal extension in the northwesternMorris et al., 2000). Second, the origins of magmaticUSA. Evidence supporting this contention comes fromrocks in the AVP are poorly understood because thereabundant isotopic ages of metamorphic and igneous rocksexist few detailed studies aimed at deciphering the petro-exposed in the cores of extensional complexes in Idahologic evolution of magmas erupted from individual erupt-and Washington State that indicate tectonic elevationive centers. This lack of detailed studies has led toand denudation at >45–50 Ma (Burchfield et al., 1992),uncertainty on a number of issues ranging from the roleand interbedded tuffs in synextensional basin-fill depositsof crustal interaction during differentiation of the calc-that range from 46 Ma to younger than 30 Ma ( Janecke &alkaline magmas to the significance of the across-strikeSnee, 1993). Nevertheless, although temporally associatedK2O enrichments in the AVP.with regional crustal extension, there is little evidenceTo address some of these issues and provide a referencewithin the AVP for the presence of major early Tertiary
point for future studies we present results from ourtectonic extensional faults. Several large-displacement
investigation of Washburn volcano, one of the principal extensional structures involving the volcanic rocks areeruptive centers for calc-alkaline magmas in the AVP present, but these are generally regarded as featuresand the type locality for the Washburn Group of the related to east-directed gravitational sliding off the grow-Absaroka Volcanic Supergroup of Smedes & Prostka ing volcanic highland (Hauge, 1985).(1972; Fig. 1). Overall, this work represents the first At present, the AVP covers >23 000 km2 in north-detailed study of a calc-alkaline (sensu stricto) eruptive western Wyoming and southwestern Montana (Fig. 1).center in the AVP. The goals of this paper are (1) to It has been tentatively correlated with other Eocenedocument the ages and ranges of major element, trace volcanic rocks in southwestern Montana, the Challiselement, and isotopic (Sr and Nd) compositions of volcanic rocks of Idaho, and the Colville igneous complexmagmatic products at Washburn volcano, (2) to place of NE Washington, although these fields are not linkedconstraints from phenocryst compositions on the crys- by outcrops (Fig. 1; Armstrong, 1978; Dudas, 1991).tallization history of Washburn rocks, (3) to evaluate Undoubtedly, much of the original extent of the volcanicthe possible roles of different petrologic processes in field is eroded or covered by Miocene and youngerproducing the spectrum of compositions observed, and (4) volcanic rocks of the Snake River Plain–Yellowstoneto use the information in (1)–(3) to develop a generalized Plateau fields, and continuity with volcanic rocks ofmodel for the evolution of calc-alkaline magmas at similar age may be obscured. However, apparent regionalWashburn volcano and in the AVP. geochemical trends within the AVP, such as increasing
The origin of intermediate composition rocks at the K2O contents eastward, suggest it may be separate.volcano is best explained by mixing between variably According to correlations advanced by Smedes & Pro-fractionated and contaminated mafic magmas and stka (1972) for Eocene rocks exposed in Yellowstoneheterogeneous silicic partial melts of Archean granulite- National Park, three groups make up the Absarokafacies rocks in the deep crust. Despite chemical evidence Volcanic Supergroup. In ascending stratigraphic orderfor mixing, there is very little textural or mineral chemistry these are (1) the Washburn Group, (2) the Sunlightevidence in the hybrid magmas to support the model. A Group, and (3) the Thorofare Creek Group. The arealscenario where hybrid magmas produced in the deep distribution of the three groups, illustrated in Fig. 1, incrust ascend to shallow reservoirs and crystallize low- part reflects the prevalent thinking at the time of thepressure mineral assemblages dominated by plagioclase study of Smedes & Prostka (1972) that volcanic activityis preferred to explain these relationships. These results migrated SE along two subparallel belts of intrusive–have important implications for the interpretation of extrusive centers: the ‘K-poor’ (calc-alkaline) western beltigneous textures and their bearing on rock-forming pro- and the ‘K-rich’ (shoshonitic) eastern belt (Chadwick,cesses and the significance of across-strike geochemical 1970). However, subsequent stratigraphic and geo-
chronologic studies in eastern and southern areas of thevariations in the AVP.
664
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Fig. 1. Map of the Absaroka Volcanic Province showing the stratigraphic units of Smedes & Prostka (1972). Black areas represent locations ofprincipal vent complexes and intrusive centers discussed by Chadwick (1970). Thick dashed line shows the approximate division between westernK-poor (calc-alkaline) and eastern K-rich (shoshonitic) belts (after Chadwick, 1970). Single dot-dashed line is the boundary of YellowstoneNational Park. Inset shows the locations of early-to-middle Eocene magmatic fields (after Chadwick, 1985; Holder & Holder, 1988; Dudas,1991; Norman & Mertzman, 1991; Wheeler et al., 1991; Luedke, 1994). Numbers refer to: 1, Sloko Volcanic Province; 2, Francois Lake igneouscomplex; 3, Colville igneous complex; 4, Clarno volcanics; 5, Challis Volcanic Province; 6, Absaroka Volcanic Province; 7, Montana AlkalicProvince; 8, Black Hills. Diagonally ruled field shows inferred extent of Archean cratonic Wyoming Province (Dutch & Nielsen, 1990).
volcanic sequence are incompatible with the no- of these complications we illustrate the inferred locationof the K2O dividing line of Chadwick (1970) and themenclature established in Yellowstone National Park (e.g.
Brown, 1982; Eaton, 1982; Sundell & Eaton, 1982; stratigraphic subdivisions of Smedes & Prostka (1972) inFig. 1 because this facilitates comparison with previousHague, 1985; Decker, 1990; Hiza, 1999). Even within
the park itself, Smedes & Prostka (1972) encountered studies. Our work also demonstrates that mafic magmaserupted at Washburn volcano and the Electric Peak–major difficulties in establishing regional stratigraphic
subdivisions because many of their formational units Sepulcher Mountain eruptive center (Lindsay & Feeley,1999, and this study) are not especially potassium rich.consist of monotonous sequences of andesitic vol-
caniclastic rocks, which are difficult to distinguish at We recognize, however, that the stratigraphy and eruptivehistory of the AVP are undoubtedly more complex thandifferent localities. Furthermore, Hiza (1999) has recently
argued that mafic magmas (i.e. <56 wt % SiO2) are originally envisioned (e.g. Hiza, 1999).Seismic refraction studies indicate that the crust be-characteristically potassic throughout the AVP and that
if only these compositions are considered, little regular neath the AVP at present is>45–50 km thick (Prodehl &Lipman, 1989). Exposed basement rocks include Archeangeographic variation in magma chemistry exists. In spite
665
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
crystalline rocks of the Wyoming Province (Fig. 1), which The samples examined in this study were chosenaccording to the stratigraphic units of Prostka et al. (1975;are mainly granitoid gneisses that intruded high-gradeFig 2) from well-exposed, vertical stratigraphic sectionsmetasedimentary and metavolcanic rocks at >2·8 Ga,dominated by vent-facies rocks. Temporal and spatialand shallow marine carbonate and clastic sedimentaryvariations in bulk chemistry for lava flows within therocks ranging in age from Cambrian to CretaceousWashburn volcanic sequence are shown schematically in(Ruppel, 1972; Wooden & Mueller, 1988). Deep- to mid-Fig. 3. In Fig. 3 (and subsequent figures) we designatecrustal lithologies are represented by mafic to silicicwith different symbols lava flows exposed on Mt. Wash-Archean granulite-facies xenoliths carried to the surfaceburn and those in the SW Washburn Range to the westby Eocene alkalic magmas in the Crazy, Bearpaw, andof the Grand Loop Road because these have differentHighwood Mountains (Dudas et al., 1987; Collerson etcompositional ranges. Lava flows and dikes in the SWal., 1989; Joswiak, 1992), and late Cenozoic basalticWashburn Range consist of a crudely bimodal packagemagmas of the Snake River Plain (Leeman et al., 1985).of olivine+ pyroxene basaltic andesites and amphibole-Many of these xenoliths are similar to granulite-faciesbearing dacites, whereas dikes, stratigraphically higherrocks worldwide in that they have relatively low contentslava flows, and the Sulphur Creek stock to the east andof Rb, U, and heavy rare earth elements (HREE), al-NE on Mt. Washburn are predominantly olivine +though they are light rare earth element (LREE) enrichedpyroxene basaltic andesites and pyroxene ± amphibole(Leeman et al., 1985; Joswiak, 1992).andesites. Included in this latter sequence are the stra-tigraphically highest exposed lava flows on Mt. Washburnthat Prostka et al. (1975) designated as part of the Langford
Stratigraphy Formation of the Thorofare Creek Group of Smedes &Washburn volcano is a major calc-alkaline eruptive center Prostka (1972; Figs 2 and 3). Although these flows werein the AVP and is the largest Eocene volcanic center originally interpreted as younger (middle to upper Eo-exposed in Yellowstone National Park. It is the primary cene) and erupted from vents much more distal thansource area for the Lamar River Formation, the eastern other units at the volcano, our work shows that they aremember of the Washburn Group of Smedes & Prostka comparable in age and composition with other flows on(1972). The Lamar River Formation is particularly well Mt. Washburn. We therefore consider all exposed unitsknown to visitors of Yellowstone National Park because to be derived from the same or very similar magmaticin it are preserved the famous upright fossil forests (Dorf, systems.1964). Washburn volcano has been previously mappedby Schultz (1962; 1:30 000) and Prostka et al. (1975;1:62 500; Fig. 2).
GEOCHRONOLOGYThe eroded northern flank of Washburn volcano inthe vicinity of Mt. Washburn, and Hedges and Dunraven To ascertain the timing of magmatism at WashburnPeaks in the SW Washburn Range, consists of>1300 m volcano we determined 40Ar/39Ar ages from phenocrystof volcanic vent facies strata, mainly of the Lamar River and groundmass samples of the stratigraphically highestFormation, that include dikes, lava flows, flow breccias, and lowest exposed lava flows together with a biotiteand debris flow deposits that dip up to 30° away from separate from the Sulphur Creek stock (Table 1). Thethe primary vent region (Fig. 2; Prostka et al., 1975). data are graphically presented in the form of 40Ar/39ArThe lava flows and dikes are largely pyroxene basaltic age spectra and 39Ar/40Ar vs 36Ar/40Ar isochron diagramsandesites and andesites, although numerous amphibole- in Fig. 4. All errors on ages and intercepts reported inbearing dacitic lava flows are present near the base Fig. 4 are 2�. The errors on individual steps, graphicallyof the sequence. With increasing distance from Mt. represented by the width of rectangular boxes on theWashburn and Hedges and Dunraven Peaks the vent- age spectrum diagrams, also represent a 2� level offacies rocks grade into alluvial-facies lithologies consisting confidence. The sample localities are shown in Fig. 2.of epiclastic volcanic conglomerate and breccia, volcanic Details of the analytical procedures are described in thesandstone and siltstone, and ashfall tuff deposits. The Appendix and a summary of the results is presented insouthern flank of Washburn volcano is truncated by Table 1. The full dataset may be downloaded from thethe northern segment of the Yellowstone Caldera fault, Journal of Petrology Web site at http://www.petrology.exposing the interior of the volcano (Fig. 2). Here, oupjournals.org.fine-grained biotite tonalite of the Sulphur Creek stock Plagioclase and amphibole phenocrysts and a sampleintrudes stratigraphically low Lamar River Formation of fine-grained groundmass were separated from a daciticvolcanic rocks. This stock is similar in composition and lava flow at the base of Hedges Peak (MW971), theage (see below) to the Eocene volcanic rocks and rep- stratigraphically lowest eruptive unit, to date the initiation
of volcanic activity at Washburn volcano. Theoretically,resents a shallow intrusion related to Washburn volcano.
666
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Fig
.2.
Sim
plifi
edge
olog
icm
apof
Was
hbur
nvo
lcan
oan
dsu
rrou
ndin
gar
ea(m
odifi
edfr
omPr
ostk
aet
al.,
1975
).Χ
,lo
catio
nsof
sam
ples
anal
yzed
inth
isst
udy.
Inse
tsh
ows
the
loca
tion
ofW
ashb
urn
volc
ano
inY
ello
wst
one
Nat
iona
lPar
k.
667
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 3. Schematic composite stratigraphic column for Washburn volcano combining total thicknesses of strata from the SW Washburn Rangeand Mt. Washburn areas. Lava flows are indicated by solid patterns; clastic units are indicated by stipple patterns (excluding brecciated autoclasticlava flow tops and bases). Geochemical data panels show the compositional variations of magmas with stratigraphic position. Open symbols arefor samples from the SW Washburn Range and filled symbols are for samples from Mt. Washburn. Circles, squares, and triangles are for basalticandesitic, andesitic, and dacitic composition rocks, respectively. Note: (1) reinterpretation of ‘Langford Formation’ flows on Mt. Washburn aslate Washburn volcano units based on data presented in this study; (2) bimodal assemblage of dacitic and basaltic andesitic lavas in lower partof section from SW Washburn Range; (3) dominantly andesitic lavas in upper part of section from Mt. Washburn.
Table 1: Summary of 40Ar/39Ar results from the Washburn volcano
Sample Material Plateau age % 39ArK (steps) Isochron age 40Ar/36Ari Total fusion
(Ma) (Ma) age (Ma)
MW971 amphibole — — — — 59·3
MW971 groundmass — — 55·2±0·6 275±4 54·1
MW971 plagioclase — — — — 54·7
MW9746 biotite 53·5±0·4 78·9 (10 of 15) 52·6±0·2 323±6 61·6
MW9743 groundmass — — 51·9±0·8 258±6 44·4
All errors given are at ±2�. Analytical details described in the Appendix.
all three samples should yield identical ages because all possibly with some K-rich phyllosilicate contaminationconsistent with the high K/Ca ratio in the first step.were at high temperatures immediately before extrusion
and cooled rapidly once emplaced. The amphibole Data from the amphibole sample failed to plot on astatistically meaningful isochron, precluding an in-sample yielded an 40Ar/39Ar spectrum with evidence of
excess argon in the initial heating steps, and most of the dependent evaluation of the isotopic ratio of trappedargon and the age of the sample. An internally discordantgas released defined ages of >60 Ma (Fig. 4a). The
integrated total fusion age (roughly equivalent to a K–Ar 40Ar/39Ar age spectrum was obtained from the plagio-clase, which has an integrated fusion age of 54·7 Maage) is 59·3 Ma; however, no age plateau was obtained.
The corresponding K/Ca ratios determined from the step (Fig. 4b). Intermediate heating steps defined ages of>50 Ma, but the last 30% of gas released defined agesheating data indicate a relatively homogeneous sample,
668
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Fig. 4. (a)–(c), (e) and (g) show apparent age spectra for 40Ar–39Ar incremental heating experiments for groundmass and mineral separates fromWashburn volcano rocks. Widths of rectangular boxes indicate estimated analytical error (±2�) for each step. (d), (f ) and (h) show corresponding36Ar/40Ar vs 39Ar/40Ar isochron diagrams for the step Ar compositions measured. The isochron ages with uncertainties (indicated) are calculatedfrom the best-fitting lines through collinear step compositions following the method of York (1969).
in excess of 60 Ma. The plagioclase data also failed to plot calculated making no assumptions about the isotopiccomposition of the trapped argon, we consider this ageon a statistically meaningful isochron. The groundmass
sample also has an internally discordant 40Ar/39Ar age as the most reliable estimate for dating the emplacementof lava flow MW97-01. Moreover, the isochron age isspectrum, with a integrated fusion age of 54·06 Ma (Fig.
4c). Apart from the initial heating step, the K/Ca ratios similar to the integrated fusion ages of both the ground-mass and the plagioclase. This interpretation suggestsare consistent with argon degassing from a relatively
homogeneous sample. Isotopic data obtained from inter- that the amphibole age of >59 Ma is unreliable, and itis consistent with no published reports of such old eruptivemediate heating steps, representing >50% of the total
argon released, plot on an isochron defining a sample ages being described from the AVP. We therefore con-sider the amphibole data as anomalously old, probablyage of 55 Ma (Fig. 4d). Because the isochron age is
669
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
because of extraneous argon. A similar result was ob- porphyritic to partially glomeroporphyritic lavas andtained in the biotite data of sample MW9746 described dikes. In thin section the rocks are hypocrystalline withbelow. intersertal to pilotaxitic groundmass textures. Phenocryst
The 40Ar/39Ar age spectrum obtained for the ground- contents range from 44 to 5% (by volume) with the totalmass of sample MW9743 is internally discordant with decreasing with increasing SiO2 contents until>63 wt %an integrated fusion age of 44·4 Ma (Fig. 4e). The highly SiO2 (Fig. 5; Table 2). Dacitic rocks have widely varyingvariable K/Ca ratios are consistent with argon degassing phenocryst contents. Changes in mode with bulk com-from reservoirs of strongly varying retentivity and com- position are regular throughout the suite. Basaltic–position. When plotted on an isochron diagram, however, andesitic rocks contain plag[ cpx > ol± opx; andesiticthe data from heating steps making up >97% of the rocks contain plag > cpx > opx ± amph ± ol; andargon released define a relatively precise (MSWD = dacitic rocks contain amph + plag ± cpx ± opx3·9) age of 51·9 Ma for the effusion of this lava (Fig. 4f ). ± bio (Fig. 5). All rocks also contain Fe–Ti oxideTaken together, the isochron ages from the groundmass microphenocrysts. Glomerocrysts are common of cpx±samples of the stratigraphically lowest and highest samples Fe–Ti oxides, ol + cpx ± Fe–Ti oxides ± plagioclaseindicate that the >1 km thick accumulated Washburn ± opx, and plag + amph ± Fe–Ti oxides. Mineralvolcanic pile was constructed in >3 my. inclusion patterns and the occurrence of minerals in
Biotite from the Sulphur Creek stock sample MW9746 the glomerocrysts, together with textural features andyielded an 40Ar/39Ar age spectrum with evidence of compositions of the phenocrysts described below, suggestextraneous argon in the first few heating steps with an the following generalized crystallization sequences. Ba-integrated fusion age of 61·6 Ma and a well-defined age saltic andesitic magmas crystallized olivine followed byplateau of 53·5± 0·4 Ma (Fig. 4g). However, an isochron Fe–Ti oxides, clinopyroxene + plagioclase, and ortho-with these data (MSWD = 3·2) indicates a slightly pyroxene. Andesitic magmas crystallized plagioclase fol-younger age of 52·6 ± 0·2 Ma (Fig. 4h). The isochron lowed by clinopyroxene+ orthopyroxene, Fe–Ti oxidesage of 52·6 Ma is our preferred date for the sample, and and then amphibole in some cases. Dacitic magmas,the 40Ar/36Ar ratio of the trapped argon component in which pyroxenes are rare or absent, precipitated(323) is significantly greater than present-day atmosphere plagioclase followed by Fe–Ti oxides, amphibole, and(295·5) and is consistent with ‘excess’ argon as previously then biotite when present. Groundmass assemblages in-inferred for the amphibole in sample MW971. clude glass (partially to pervasively devitrified), plagio-
On the basis of the 40Ar/39Ar data presented here clase, clinopyroxene, and orthopyroxene. Zircon andwe interpret magmatism at Washburn volcano to have apatite are common accessory phases in the intermediatecommenced as early as 55 Ma and possibly continued and silicic lavas. Xenocrysts, identified by non-equi-until at least 52 Ma. These ages bracket the 53·4 ± 0·3 librium compositions (see below) or magmatic reactionage reported by Hiza (1999) for a dacitic block and ash textures, are present but rare. Furthermore, we identifiedflow at the base of Sepulcher Mountain, located>50 km no clear petrographic evidence for mixing or minglingto the NW of Mt. Washburn (Fig. 1). The sequence of between compositionally disparate magmas, such as therocks exposed on Sepulcher Mountain represents the presence of undercooled blobs of mafic magma that aretype section for the Sepulcher Formation, the western frequently found in many andesitic to dacitic rocksmember of the Washburn Group of Smedes & Prostka (Bacon, 1986; Wilcox, 1999).(1972). Because these rocks are compositionally and The Sulphur Creek stock is tonalitic to quartz dioriticpetrographically identical to rocks exposed at Washburn with plag > qtz > bio > cpx = opx > Fe–Ti oxide >volcano (Lindsay & Feeley, 1999), we concur with the amph. Texturally, the stock is fine grained, phaneritic,opinion of Smedes & Prostka (1972) that rocks within and subophitic in that late-crystallizing anhedral quartzthe Sepulcher and Lamar River Formations were erupted partially encloses elongate plagioclase and biotite grains.from similar and nearly contemporaneously active vol-canic centers. In addition, our results are also consistentwith the suggestion of Hiza (1999) that the oldest rocks
Olivinein the AVP are calc-alkaline lavas at present exposed inthe northwestern part of the field. Olivine generally occurs as equant, euhedral to subhedral
phenocrysts in basaltic andesitic lavas. In the majority ofthese samples the phenocrysts range in size from about4·0 to 0·5 mm, although populations in individual samples
PETROGRAPHY AND SILICATE typically have a much narrower range. In a few samplesMINERAL CHEMISTRY the size variation is continuous from a maximum of
1·0 mm to a minimum of 0·1 mm. Additionally, a fewWashburn igneous rocks investigated in this study aregenerally non- to slightly vesicular (<3 vol. % vesicles), andesitic and dacitic rocks contain small amounts of
670
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Tab
le2
:M
ajor
elem
ent,
trac
eel
emen
t,Sr
and
Nd
isot
opic
ratios
,an
dm
odal
data
for
Was
hbur
nvo
lcan
oig
neou
sro
cks
SW
Was
hb
urn
Ran
ge
Sam
ple
:M
W97
13M
W97
17M
W97
12M
W97
16M
W97
7O
P98
73M
W97
2M
W97
14M
W97
15M
W97
9O
P98
61O
P98
74O
P98
75M
W97
6M
W97
11O
P98
82M
W97
8
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eLa
vaLa
vaLa
vaLa
vaLa
vaLa
vaD
ike
Lava
Cla
ssifi
cati
on
:B
AB
AB
AB
AB
AB
AB
AB
AB
AB
AA
AA
AD
DD
XR
Fw
t%
SiO
252
·39
54·3
254
·52
54·5
955
·13
55·1
455
·62
55·6
256
·46
56·9
157
·10
57·4
557
·94
62·6
364
·09
64·1
764
·29
Al 2
O3
14·9
814
·09
15·4
914
·16
15·3
114
·51
14·6
014
·07
17·4
414
·50
17·4
914
·97
15·1
116
·18
16·2
115
·91
15·6
9
Fe2O
3T9·
989·
078·
878·
888·
758·
288·
288·
517·
687·
487·
366·
997·
004·
854·
824·
234·
52
TiO
20·
900·
750·
820·
750·
770·
770·
750·
700·
820·
690·
810·
680·
670·
530·
510·
430·
42
Mn
O0·
160·
140·
140·
150·
130·
140·
130·
140·
110·
170·
100·
120·
100·
040·
080·
040·
05
CaO
8·58
8·69
8·33
8·79
8·46
8·74
7·95
7·83
7·30
7·06
6·92
6·78
6·54
3·19
4·61
4·56
3·85
Mg
O8·
658·
216·
278·
056·
357·
087·
367·
083·
877·
793·
936·
936·
182·
333·
032·
862·
98
K2O
1·25
1·53
1·74
1·41
1·58
1·52
1·72
1·67
2·11
1·75
1·87
1·86
1·97
4·96
1·88
2·17
4·39
Na 2
O2·
582·
472·
622·
622·
712·
732·
872·
973·
292·
974·
013·
383·
362·
973·
854·
123·
23
P2O
50·
180·
230·
210·
230·
220·
230·
230·
230·
220·
200·
230·
200·
200·
210·
190·
180·
17
Tota
l∗99
·65
99·5
099
·00
99·6
299
·41
99·1
499
·50
98·8
299
·31
99·5
299
·81
99·3
699
·07
97·8
999
·27
98·6
799
·60
XR
Fp
pm
V17
919
918
418
518
520
118
816
617
915
516
814
614
989
7978
80
Cr
415
420
206
421
497
422
324
507
3547
133
355
312
6311
514
010
4
Ni
134
114
6311
758
116
9615
813
174
1413
610
730
5363
48
Sc
2625
2826
2426
2123
2118
2221
188
915
10
Ba
589
748
941
709
917
738
945
1008
1061
1063
1219
1088
1137
1585
1372
1356
2013
Rb
2227
3225
4128
2727
3029
2930
3581
7938
55
Sr
463
623
642
599
652
620
676
694
748
648
757
668
675
535
626
800
749
Zr
104
109
126
108
116
109
124
121
140
128
141
132
135
162
176
142
132
Y17
1516
1414
1614
1315
1316
1314
1413
109
Nb
54
55
45
54
55
67
66
56
6
Pb
68
147
129
1112
137
1414
1318
1721
19
Th
63
33
41
53
55
52
33
53
4
671
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Tab
le2
:co
ntin
ued
SW
Was
hb
urn
Ran
ge
Sam
ple
:M
W97
13M
W97
17M
W97
12M
W97
16M
W97
7O
P98
73M
W97
2M
W97
14M
W97
15M
W97
9O
P98
61O
P98
74O
P98
75M
W97
6M
W97
11O
P98
82M
W97
8
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eLa
vaLa
vaLa
vaLa
vaLa
vaLa
vaD
ike
Lava
Cla
ssifi
cati
on
:B
AB
AB
AB
AB
AB
AB
AB
AB
AB
AA
AA
AD
DD
INA
Ap
pm
Sc
28·0
23·3
18·7
19·0
10·3
Cs
0·7
0·9
0·7
0·8
2·4
La15
·724
·333
3039
Ce
31·0
46·7
57·0
51·5
69
Nd
14·6
19·0
21·0
17·0
23
Sm
3·20
3·68
4·8
4·2
4·5
Eu
1·19
1·23
1·42
1·21
1·32
Tb
0·56
0·51
0·46
0·39
0·38
Yb
1·60
1·50
1·5
1·7
1·2
Lu0·
220·
190·
240·
220·
21
Hf
2·94
2·82
3·60
3·40
5·0
Ta0·
250·
210·
27—
—
U0·
730·
77—
0·40
2
(87S
r/86
Sr)
m0·
7058
470·
7057
27
(143 N
d/14
4 Nd
) m0·
5118
900·
5116
32
(87S
r/86
Sr)
i†0·
7057
440·
7056
42
(143 N
d/14
4 Nd
) i0·
5118
440·
5115
91
Mo
dal
ph
eno
crys
ts(v
ol.
%)
ol
9·2
8·6
6·1
5·2
2·0
8·3
1·8
7·6
0·4
4·4
2·6
—0·
2—
0·3
——
pla
g21
·220
·027
·420
·613
·69·
321
·610
·425
·59·
711
·66·
412
·812
·11·
54·
32·
3
cpx
2·9
10·4
5·7
13·6
7·6
14·2
10·1
12·9
6·0
10·1
2·1
8·1
13·9
8·5
——
—
op
x0·
22·
42·
51·
20·
61·
53·
41·
30·
85·
30·
56·
13·
7—
3·1
——
oxi
de
0·9
1·3
1·6
1·5
0·2
—0·
40·
71·
40·
41·
60·
30·
81·
80·
71·
00·
2
amp
hib
ole
——
——
——
——
——
——
—3·
61·
99·
512
·6
bio
——
——
——
——
——
——
——
——
—
qtz
——
——
——
——
——
——
——
——
—
Tota
l34
·442
·743
·442
·123
·933
·337
·432
·934
·130
·018
·420
·931
·426
·17·
614
·815
·0
672
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
SW
Was
hb
urn
Ran
ge
Mt.
Was
hb
urn
Sam
ple
:M
W97
10M
W97
5O
P98
83M
W97
1M
W97
3O
P98
84O
P98
81M
W97
4M
W97
24M
W97
30M
W97
27M
W97
28M
W97
35M
W92
0M
W97
34M
W97
23M
W97
25
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eD
ike
Dik
eLa
vaD
ike
Lava
Dik
eLa
vaD
ike
Lava
Cla
ssifi
cati
on
:D
DD
DD
DD
DB
AB
AB
AB
AB
AB
AA
AA
XR
Fw
t%
SiO
264
·31
64·6
564
·85
65·0
165
·32
66·3
866
·65
67·1
154
·79
55·1
155
·69
55·9
556
·40
56·9
257
·36
57·4
857
·85
Al 2
O3
15·9
816
·19
15·4
315
·67
16·0
116
·01
15·1
415
·31
14·8
916
·32
15·9
316
·06
16·1
715
·82
16·5
315
·76
16·4
1
Fe2O
3T4·
494·
564·
044·
405·
113·
342·
633·
488·
677·
788·
527·
977·
788·
077·
176·
977·
39
TiO
20·
420·
500·
450·
420·
490·
440·
330·
430·
780·
840·
790·
720·
790·
790·
780·
740·
76
Mn
O0·
090·
070·
040·
050·
020·
050·
060·
050·
140·
120·
140·
110·
120·
120·
120·
110·
11
CaO
4·03
4·42
4·74
4·37
3·70
3·22
3·57
3·45
8·81
7·43
8·44
7·94
8·02
8·17
7·25
7·46
6·80
Mg
O2·
862·
633·
153·
042·
081·
742·
021·
616·
434·
174·
695·
225·
284·
464·
925·
424·
64
K2O
2·20
2·15
2·27
2·14
3·02
3·11
2·32
2·61
1·77
1·96
1·86
1·94
1·71
1·53
1·89
1·86
2·27
Na 2
O4·
353·
693·
973·
873·
503·
884·
473·
752·
603·
092·
933·
133·
322·
893·
263·
043·
06
P2O
50·
180·
200·
170·
170·
190·
160·
130·
150·
230·
270·
240·
230·
240·
220·
220·
200·
24
Tota
l98
·91
99·0
699
·11
99·1
499
·45
98·3
297
·32
97·9
499
·11
97·0
999
·23
99·2
799
·83
98·9
999
·49
99·0
499
·52
XR
Fp
pm
V71
8581
7674
6355
5220
318
018
617
418
216
517
116
415
6
Cr
114
7915
912
771
6275
4222
454
103
143
120
289
8417
010
4
Ni
4836
6254
3231
4125
5116
3550
4270
3156
42
Sc
78
1111
79
78
2419
2425
2226
2320
20
Ba
1530
1349
1372
1333
1417
1547
1579
1445
787
1268
943
1049
1013
1182
1003
1140
1162
Rb
4570
3733
6164
4281
2833
3335
2637
3634
39
Sr
782
583
752
800
527
464
772
510
624
841
672
737
705
693
632
664
771
Zr
142
183
134
141
185
172
145
180
112
137
121
133
123
124
126
131
144
Y10
139
913
127
1316
1615
1414
1714
1516
Nb
55
64
65
56
44
45
44
55
6
Pb
2115
1819
1719
2522
913
1011
1112
1414
15
Th
35
11
66
15
26
26
44
52
3
673
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Tab
le2
:co
ntin
ued
SW
Was
hb
urn
Ran
ge
Mt.
Was
hb
urn
Sam
ple
:M
W97
10M
W97
5O
P98
83M
W97
1M
W97
3O
P98
84O
P98
81M
W97
4M
W97
24M
W97
30M
W97
27M
W97
28M
W97
35M
W92
0M
W97
34M
W97
23M
W97
25
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eD
ike
Dik
eLa
vaD
ike
Lava
Dik
eLa
vaD
ike
Lava
Cla
ssifi
cati
on
:D
DD
DD
DD
DB
AB
AB
AB
AB
AB
AA
AA
INA
Ap
pm
Sc
9·2
8·5
6·8
21·0
25·2
Cs
2·1
0·8
1·1
1·2
1·1
La30
4137
3425
Ce
51·3
6161
·467
·445
·7
Nd
14·6
3125
·728
·316
Sm
2·82
4·5
3·58
4·87
4·4
Eu
0·89
0·9
1·10
1·56
1·42
Tb
0·28
—0·
360·
630·
48
Yb
1·10
1·7
1·01
1·83
1·7
Lu0·
150·
210·
150·
270·
23
Hf
3·20
43·
653·
163·
1
Ta—
0·9
0·42
0·30
0·26
U0·
971·
170·
770·
2
(87S
r/86
Sr)
m0·
7063
150·
7082
630·
7059
64
(143 N
d/14
4 Nd
) m0·
5114
170·
5111
810·
5114
02
(87S
r/86
Sr)
i0·
7061
900·
7079
170·
7058
79
(143 N
d/14
4 Nd
) i0·
5113
760·
5111
520·
5113
66
Mo
dal
ph
eno
crys
ts(v
ol.
%)
ol
——
——
——
——
3·0
1·5
4·1
2·7
5·0
6·4
1·8
1·2
—
pla
g5·
610
·17·
26·
118
·08·
89·
815
·924
·67·
830
·613
·416
·87·
511
·013
·015
·3
cpx
——
0·7
——
—0·
5—
13·5
4·0
7·2
6·9
10·7
7·6
4·6
5·6
2·2
op
x—
3·6
——
——
—1·
70·
90·
70·
70·
40·
20·
71·
14·
72·
0
oxi
de
1·3
1·8
0·3
0·4
1·8
0·9
0·4
1·2
0·4
3·6
0·2
0·8
0·5
0·3
0·4
0·3
1·8
amp
hib
ole
8·9
1·6
6·9
13·3
6·4
4·1
2·1
2·4
——
——
——
——
—
bio
——
——
——
0·8
——
——
——
——
——
qtz
——
——
——
——
——
——
——
——
—
Tota
l15
·817
·215
·119
·826
·113
·813
·621
·342
·317
·642
·824
·233
·122
·518
·924
·821
·3
674
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Mt.
Was
hb
urn
Su
lph
ur
Cre
ekS
tock
Sam
ple
:M
W97
31M
W97
21M
W97
19M
W97
32M
W92
9M
W97
41M
W97
43M
W97
44M
W97
39M
W97
45M
W97
38M
W97
46
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eLa
vaLa
vaS
tock
Cla
ssifi
cati
on
:A
AA
AA
AA
AD
DD
A
XR
Fw
t%
SiO
258
·94
59·1
859
·32
59·7
861
·04
61·6
061
·74
61·8
663
·02
63·5
564
·33
59·8
1
Al 2
O3
17·5
215
·89
17·8
616
·01
14·4
214
·36
16·2
116
·13
17·1
716
·94
17·3
316
·63
Fe2O
3T6·
976·
567·
036·
736·
125·
095·
055·
575·
104·
414·
586·
00
TiO
20·
740·
700·
740·
670·
540·
460·
540·
570·
600·
500·
490·
71
Mn
O0·
090·
130·
110·
140·
080·
080·
080·
100·
050·
070·
050·
09
CaO
6·08
6·73
5·78
6·35
5·27
4·79
5·40
5·50
5·14
4·78
4·66
6·12
Mg
O2·
924·
672·
534·
236·
085·
943·
423·
971·
792·
351·
324·
01
K2O
2·11
2·39
2·16
1·89
2·08
2·62
2·63
2·15
2·27
2·95
2·47
2·06
Na 2
O3·
922·
883·
833·
683·
503·
463·
473·
974·
043·
694·
183·
82
P2O
50·
320·
200·
250·
200·
170·
160·
230·
200·
210·
240·
220·
23
Tota
l99
·60
99·3
299
·62
99·6
899
·31
98·5
698
·77
100·
0299
·39
99·4
999
·61
99·4
8
XR
Fp
pm
V14
115
714
515
311
888
105
110
102
7980
153
Cr
713
811
121
360
402
9914
212
408
82
Ni
337
1146
157
235
4160
817
637
Sc
1419
1519
1613
1318
813
1121
Ba
1453
1306
1320
1139
1286
1329
1640
1279
1338
1759
1460
1245
Rb
3844
4035
3833
4645
4651
5538
Sr
953
706
748
741
694
817
1025
785
788
962
826
799
Zr
158
138
152
131
134
124
160
140
148
171
153
145
Y14
1515
1311
912
1213
1412
14
Nb
85
65
54
65
58
55
Pb
1613
1415
1519
2418
2123
1715
Th
51
64
41
54
55
55
675
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Tab
le2
:co
ntin
ued
Mt.
Was
hb
urn
Su
lph
ur
Cre
ekS
tock
Sam
ple
:M
W97
31M
W97
21M
W97
19M
W97
32M
W92
9M
W97
41M
W97
43M
W97
44M
W97
39M
W97
45M
W97
38M
W97
46
Ro
ckty
pe:
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Lava
Dik
eLa
vaLa
vaS
tock
Cla
ssifi
cati
on
:A
AA
AA
AA
AD
DD
A
INA
Ap
pm
Sc
14·2
13·5
15·4
12·0
8·9
6·4
17·6
Cs
1·2
1·6
0·6
1·4
1·4
1·8
1·0
La47
·536
33·7
40·1
5636
33·0
Ce
90·8
6359
·467
·297
6061
·6
Nd
30·9
2218
·027
·733
2129
·0
Sm
5·49
4·8
3·48
3·92
5·5
4·1
4·64
Eu
1·64
1·48
1·16
1·27
1·54
1·19
1·59
Tb
0·46
—0·
310·
370·
410·
330·
65
Yb
1·30
1·7
1·04
1·05
1·2
1·2
1·80
Lu0·
140·
220·
140·
170·
170·
180·
18
Hf
3·70
3·9
3·28
2·57
4·3
3·9
2·62
Ta—
—0·
260·
330·
19—
0·25
U1·
471
0·84
1·43
22
0·90
(87S
r/86
Sr)
m0·
7063
070·
7071
140·
7061
420·
7063
59
(143 N
d/14
4 Nd
) m0·
5114
750·
5114
030·
5114
250·
5115
12
(87S
r/86
Sr)
i0·
7062
200·
7069
950·
7060
440·
7062
56
(143 N
d/14
4 Nd
) i0·
5114
380·
5113
620·
5113
950·
5114
78
Mo
dal
ph
eno
crys
ts(v
ol.
%)
ol
——
——
—0·
93·
3—
——
——
pla
g16
·310
·425
·312
·713
·69·
415
·013
·52·
617
·62·
748
·1
cpx
1·9
6·8
0·8
6·2
8·4
1·7
4·0
1·8
—0·
7—
9·8
op
x0·
74·
51·
13·
76·
41·
94·
93·
4—
2·0
—3·
4
oxi
de
3·1
1·8
1·7
2·5
1·2
0·3
0·4
1·3
0·7
0·9
0·8
3·7
amp
hib
ole
0·5
2·3
—1·
50·
50·
90·
9—
0·9
2·1
0·5
1·1
bio
——
——
——
——
——
—8·
3
qtz
——
——
——
——
——
—25
·6
Tota
l22
·525
·829
·026
·630
·017
·525
·220
·24·
223
·24·
110
0·0
Tota
lir
on
rep
ort
edas
Fe2O
3.C
lass
ifica
tio
n:
BA
,b
asal
tic
and
esit
e;A
,an
des
ite;
D,
dac
ite.
∗Wei
gh
tlo
sso
nig
nit
ion
no
td
eter
min
ed.
†In
itia
lis
oto
pic
rati
os
calc
ula
ted
at53
Ma.
676
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Table 3: Representative olivine analyses
Sample: MW9713 MW9717 MW9730
core rim core rim core rim
SiO2 38·90 38·29 39·74 39·09 38·66 37·69
TiO2 0·00 0·00 0·02 0·01 0·01 0·04
Al2O3 0·01 0·03 0·01 0·02 0·02 0·02
Cr2O3 0·02 0·03 0·06 0·01 0·00 0·01
FeO∗ 15·46 20·43 14·47 18·05 19·07 21·59
MnO 0·18 0·31 0·22 0·27 0·29 0·35
MgO 44·91 40·55 45·67 42·55 42·02 40·01
CaO 0·20 0·29 0·13 0·17 0·16 0·16
NiO 0·19 0·11 0·30 0·14 0·04 0·08
Total 99·88 100·04 100·62 100·30 100·28 99·94
Fo 84 78 85 81 80 77
(Table 3). Fe2+/Mg ratios of olivine phenocryst coresare plotted versus those of three whole rocks in Fig. 6a.In this figure the whole-rock Fe2O3/FeO weight ratio isassumed to be >0·24 based on the Kilinc et al. (1983)expression for Fe speciation, an f (O2) of nickel–nickeloxide (NNO; Huebner & Sato, 1970), and a temperatureof 1000°C (see below). In general, minimum olivine coreFe2+/Mg ratios increase with increasing rock Fe2+/Mgratios and appear to indicate equilibrium between theolivine and the bulk rock based on a KD (=[Fe/Mg]ol/[Fe/Mg]rock) = 0·3 ± 0·03 (Roeder & Emslie, 1970;Wagner et al., 1995).
PyroxeneClinopyroxene occurs in Washburn volcanic rocks as
Fig. 5. Variation of modal abundances of phenocryst phases in Wash- isolated, euhedral to subhedral phenocrysts up to 6 mmburn lavas and dikes (Sulphur Creek Stock excluded) vs wt % SiO2. across, as compact, often rounded glomerocrystic ag-Symbols as in Fig. 3. Large symbols in pyroxene panel represent gregates up to 1 cm in diameter, as a microphenocrystclinopyroxene; small symbols represent orthopyroxene. All rocks also
phase in the groundmass, and as subhedral to euhedralcontain microphenocrysts of Fe–Ti oxides and a few dacitic rockscontain small amounts of biotite phenocrysts. rims surrounding orthopyroxene cores in basaltic andes-
itic rocks. In dacitic rocks clinopyroxene phenocrysts maybe surrounded by corona of amphibole. Orthopyroxene
olivine (Fig. 5). Although we did not analyze any of these occurs as: (1) individual subhedral to euhedral lath-grains, they are probably xenocrysts or relicts because shaped phenocrysts 0·5–1·0 mm across; (2) as crystals ingrain boundaries are ragged, variably resorbed, and monomineralogic clots and the glomerocrysts describedhave corona of clinopyroxene or orthopyroxene plus above; (3) as smaller crystals together with plagioclase andplagioclase. In nearly all rocks olivine is partially altered to Fe–Ti oxides in reaction coronas surrounding amphibolebowlingite, particularly along crystal faces and fractures. phenocrysts of some dacitic rocks; (4) as microphenocrysts
The total analyzed variation in the compositions of in the groundmass.olivine in the basaltic andesitic rocks ranges from Fo85 Pyroxene compositions are typical of calc-alkaline rockto Fo72; within single phenocrysts, rims are normally series. Clinopyroxenes (Wo34–46En43–52Fs5–16) are mainly
augite, although a few phenocrysts have endodiopsidiczonedΖ7 mol % Fo content relative to core compositions
677
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 6. (a)–(c) Fe2+/Mg in cores of olivine, clinopyroxene, and orthopyroxene phenocrysts, respectively plotted vs Fe2+/Mg in host rocks. Alldata plotted as filled symbols. Lines in diagrams represent equilibrium between minerals and whole-rock compositions. In all figures the whole-rock Fe2O3/FeO weight ratio is assumed to be >0·24 based on the Kilinc et al. (1983) expression for Fe speciation, an f (O2) of NNO (Huebner& Sato, 1970), and a temperature of 1000°C. (d) Plot of Ca/Na in cores of plagioclase phenocrysts vs Ca/Na of host rocks. Lines representexchange KD based on experimental studies discussed in text. Inset shows Ca/Na in rims of plagioclase phenocryst vs Ca/Na of host-rockgroundmass.
cores, particularly in more magnesian-rich rocks, and in this particular thin section), the relationships illustratedare consistent with an equilibrium KD (=[Fe/Mg]pyx/orthopyroxenes are bronzite–hypersthene (Wo2–5-[Fe/Mg]rock) between 0·2 and 0·3 for clinopyroxene,En67–81Fs16–30; Table 4). All Washburn clinopyroxeneswhich is similar to values (0·20–0·25) found ex-have low Al (Al/6 oxygens <0·18) and Ti (Ti/6 oxygensperimentally in 1 atm, silica-saturated and undersaturated<0·02), suggesting crystallization under low-pressure,arc lavas (Grove & Baker, 1984; Kennedy et al., 1990).shallow-crustal conditions (Fig. 7a). Pyroxene zoningEquilibrium KD values for orthopyroxene in Mg-richpatterns are generally normal. Relative to cores, mostbasaltic andesitic rocks appear higher than for clino-rims are enriched in ferrosilite content by an amountpyroxene, but probably reflect early precipitation of<9%, and they have lower contents of Cr, Al, and lessclinopyroxene followed by orthopyroxene in maficcommonly Ti. In addition, a few analyzed pyroxenemagmas. Magmatic temperatures calculated for rimsgrains are reversely zoned. The majority of these areof touching clinopyroxene–orthopyroxene pairs in threeclinopyroxene phenocrysts with rimward enrichments inlavas using the methods of Wood & Banno (1973) andMg (Ζ7% En) and less typically Ca (Ζ4% Wo) and Ti.Wells (1977) are in the range of 1012–1017°C for aReverse zoning is less common in orthopyroxene andbasaltic andesite (MW9730), 981–999°C for an andesitetypically at the limit of analytical resolution (Ζ1% En).(MW9721), and 967–968°C for a high-silica andesiteThe origin of the Mg-rich clinopyroxene rims may be(MW9743; Table 4).related to recharge and mixing with less evolved magmas
or, as suggested by the lack of reverse zonation in theWo component for many of these grains, late precipitationof Fe–Ti oxides in the groundmass and lowering of FeO
Plagioclasein the liquid.Compositions of pyroxene phenocryst cores are com- Plagioclase is the most abundant mineral in Washburn
rocks, with the exception of a few dacitic and basalticpared with whole-rock compositions in Fig. 6b and c.With the exception of an obviously xenocrystic grain in andesitic samples in which amphibole and clinopyroxene,
respectively, are slightly more abundant. In all bulk-rocka low-Mg dacite (MW974; the only cpx grain observed
678
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Tab
le4
:R
epre
sent
ativ
epy
roxe
nean
alys
es
Sam
ple
:M
W97
13M
W97
17M
W97
30M
W97
21M
W97
43M
W97
5M
W97
4C
lino
pyr
oxe
ne
core
rim
core
rim
core
rim
core
rim
core
rim
core
rim
core
rim
SiO
251
·39
50·7
151
·45
51·6
850
·19
49·6
551
·81
50·3
851
·07
50·8
4Ti
O2
0·55
0·45
0·60
0·68
0·60
0·56
0·42
0·50
0·41
0·42
Al 2
O3
2·52
2·62
2·35
2·26
1·62
3·26
2·88
2·91
2·22
2·06
Cr 2
O3
0·64
0·58
0·41
0·17
0·00
0·02
0·22
0·06
0·01
0·00
FeO∗
8·36
8·77
9·13
9·52
12·0
68·
207·
218·
3511
·69
10·3
0M
nO
0·24
0·19
0·26
0·24
0·31
0·15
0·16
0·17
0·46
0·44
Mg
O17
·23
17·0
816
·85
16·2
014
·31
15·4
416
·57
14·7
913
·40
14·0
2C
aO17
·71
17·0
717
·88
18·2
718
·73
20·5
419
·86
20·9
319
·75
20·3
2N
a 2O
0·37
0·33
0·34
0·34
0·54
0·39
0·43
0·34
0·56
0·60
Tota
l99
·21
97·9
899
·49
99·4
798
·78
98·6
899
·79
98·7
199
·84
99·3
5W
O37
·836
·938
·138
·641
·545
·542
·345
·543
·344
·9E
N51
·251
·349
·947
·644
·247
·549
·144
·740
·943
·1FS
11·0
11·8
12·1
13·7
14·3
7·0
8·6
9·8
15·7
12·0
Ran
ge1
WO
35·9
–37·
933
·6–4
0·1
38·1
–42·
737
·7–4
2·7
40·8
–44·
142
·9–4
6·3
42·3
–45·
043
·7–4
6·1
42·4
–45·
441
·2–4
4·9
EN
48·5
–51·
247
·4–5
2·2
48·5
–51·
847
·6–5
1·5
44·2
–47·
746
·4–4
7·5
44·4
–49·
143
·6–4
7·1
40·9
–50·
643
·1–4
8·0
FS10
·4–1
4·7
11·2
–15·
15·
5–12
·16·
4–13
·79·
7–14
·37·
0–10
·88·
6–11
·58·
9–10
·36·
7–15
·77·
8–12
·0O
rth
op
yro
xen
eS
iO2
52·9
053
·05
50·9
152
·20
52·4
253
·06
52·8
553
·37
53·2
552
·05
51·2
852
·07
TiO
20·
380·
370·
290·
240·
140·
180·
190·
280·
140·
190·
280·
08A
l 2O
31·
251·
332·
791·
761·
671·
291·
551·
282·
291·
583·
481·
40C
r 2O
30·
100·
120·
010·
030·
070·
000·
070·
000·
600·
020·
140·
07Fe
O∗
16·4
516
·16
15·8
615
·36
15·7
016
·65
16·8
517
·56
11·8
817
·63
15·9
417
·24
Mn
O0·
390·
400·
320·
360·
350·
420·
360·
420·
250·
430·
250·
32M
gO
25·6
525
·77
26·7
227
·23
27·5
026
·49
26·3
025
·69
29·0
425
·88
25·6
925
·98
CaO
2·36
2·27
1·42
1·37
1·19
1·30
1·18
1·23
1·36
0·99
1·66
1·02
Na 2
O0·
040·
040·
030·
050·
040·
010·
040·
060·
020·
020·
030·
00To
tal
99·7
699
·72
98·8
699
·01
99·5
299
·66
99·6
610
0·08
99·0
699
·14
99·0
598
·50
WO
4·8
4·6
3·0
2·8
2·4
2·6
2·4
2·5
2·7
2·0
3·5
2·1
EN
72·5
72·7
78·5
78·2
78·7
74·8
74·6
72·3
81·6
74·5
74·8
74·7
FS22
·722
·718
·518
·918
·922
·623
·025
·315
·723
·421
·723
·2R
ang
e1
WO
2·4–
4·9
3·0–
4·3
3·0–
3·2
2·7–
3·0
1·8–
2·8
2·4–
2·6
2·1–
2·8
2·5–
3·8
1·9–
2·5
2·0–
2·7
1·8–
4·2
1·9–
4·2
EN
72·5
–77·
267
·2–7
4·0
73·1
–78·
573
·6–7
8·2
75·4
–78·
774
·8–7
7·3
71·1
–81·
172
·3–7
7·1
71·8
–81·
672
·1–7
3·9
69·6
–74·
869
·3–7
1·0
FS20
·4–2
2·7
21·7
–29·
718
·5–2
3·2
18·9
–23·
318
·9–2
2·0
20·3
–22·
617
·5–2
6·1
20·2
–25·
315
·7–2
6·2
23·8
–25·
721
·2–2
7·1
23·2
–28·
8T
(°C
)2W
B10
1798
496
7T
(°C
)3W
1012
981
968
1 Ran
ges
are
for
all
anal
yzed
po
ints
ina
giv
ensa
mp
le.
Tem
per
atu
res
calc
ula
ted
for
rim
so
fto
uch
ing
clin
op
yro
xen
e–o
rth
op
yro
xen
ep
airs
by
the
met
ho
ds
of
2 Wo
rds
&B
ann
o(1
973)
and
3 Wel
ls(1
977)
.
679
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
With the exception of a few Ca-rich phenocrysts in thedacitic lavas, the relationships illustrated are consistentwith an equilibrium KD (=[Ca/Na]plag/[Ca/Na]rock) be-tween one and three for the andesitic and dacitic lavas.These values suggest early crystallization of most plagio-clase cores under low to moderate magmatic watercontents, based on the data of Sisson & Grove (1993),who showed that KD varies with melt water content, from1·0 at anhydrous conditions to 5·5 for melts with 6%water at 2 kbar. The more Ca-rich cores may representgrains crystallized under higher water contents, un-equilibrated crystals from higher temperature or pressurestages of evolution, xenocrysts derived from mixing withmore mafic magmas, or low-density refractory solidsretained from zones of deep-crustal melting and silicicmagma production (see Feeley & Dungan, 1996). Becausedacitic and andesitic rocks have relatively high pro-portions of these Ca-rich cores, whereas they are lessabundant in more mafic rocks (e.g. Fig. 6d), we favorthe latter hypothesis (see below). The relationships in Fig.6d for the basaltic andesitic magmas suggest equilibriumKD values around one for most plagioclase cores. Thesevalues suggest crystallization from magmas with lower
Fig. 7. (a) Stoichiometric Ti and Al per six oxygens in clinopyroxenes water contents relative to more evolved magmas orfrom Washburn volcano. Pressure fields after Stewart et al. (1996) and precipitation of plagioclase concurrently or followingreferences therein. (b) AlIV vs (Na+ K)A site occupancy for amphibole
saturation of high-Ca clinopyroxene. Also shown in Fig.phenocrysts in Washburn rocks. Shown for reference are the com-6d are equilibrium relationships between plagioclasepositional trends with temperature determined experimentally by Heltz
(1973; Kilauea and Picture Gorge) and Eggler (1972; Paracutin). phenocryst rim and liquid (i.e. groundmass) Ca/Na ratiosSymbols as in Fig. 3. determined for five rocks by electron microprobe ana-
lyses. The plagioclase phenocryst rims in these rocksstraddle or plot slightly above the 3·0 KD referencecompositions plagioclase occurs as large (0·3–4·0 mm),line. These relationships suggest equilibrium growth ofeuhedral to subhedral phenocrysts in isolation or in
glomerocrysts, and as equant microphenocrysts and lath- plagioclase rims under higher water contents than forshaped microlites in the groundmass. Phenocryst habits cores, perhaps because of residual water build-up duringare generally lath-shaped, although grains become in- crystallization of predominantly anhydrous mineral as-creasingly equant in more silica-rich rocks. There appears semblages.to be no compositional distinction between plagioclasegrains in isolation or in glomerocrysts. Compositions ofplagioclase phenocryst cores for the Washburn suite span
Amphibolefrom An83 to An40 (Table 5), although greater than two-Amphibole is present in more silica-rich Washburn rocksthirds of phenocryst cores are between An65 and An50
as phenocrysts ranging in size up to 4 mm. In many of(Fig. 8).these rocks amphibole is rimmed by opacite or lessAlthough oscillatory zoning of small magnitude is com-commonly by a fine-grained mass of plagioclase+ Fe–Timon, overall core to rim zoning patterns of plagioclaseoxides± pyroxenes, presumably as a result of breakdownphenocrysts are relatively uncomplicated; over half of allduring decompression (Foden & Green, 1992). In ad-crystals analyzed are normally zoned <15 mol % An anddition, amphibole phenocrysts in many samples are par-<20% display reverse zoning of any magnitude. Thetially altered along cleavage surfaces and cracks to a fine-maximum extent of reverse zoning observed is <12 mol %grained mass, perhaps bowlingite, and a few samplesAn and with very few exceptions it is <5 mol %. Ground-have crystals with cavernous interiors, presumably as amass crystals in all rock types have compositional rangesresult of incomplete crystal growth. In these latter crystalsthat are similar to the ranges of rim compositions andalteration can be substantial. All Washburn amphibolesthus are either identical to, or slightly more Na rich thanhave (Ca+ Na)M4 >1·34 and (Na)M4 <0·67 and thus arecore compositions (Fig. 8).calcic amphiboles according to the classification schemeThe relationships between plagioclase phenocryst cores
and whole-rock Ca/Na ratios are illustrated in Fig. 6d. of Leake (1978) and Hawthorne (1981). Amphibole
680
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Tab
le5
:R
epre
sent
ativ
epl
agio
clas
ean
alys
es
Sam
ple
:M
W97
13M
W97
17M
W97
12M
W97
30M
W97
43M
W97
21M
W97
5M
W97
4
core
rim
core
rim
core
rim
core
rim
core
rim
core
rim
core
rim
core
rim
SiO
252
·82
52·9
453
·82
52·6
953
·18
53·9
253
·66
55·9
852
·75
54·7
949
·88
50·9
754
·42
55·3
255
·53
60·8
5
Al 2
O3
28·8
028
·62
27·6
227
·84
28·5
628
·09
27·8
226
·32
28·3
627
·15
30·4
529
·60
28·7
127
·57
27·1
023
·35
FeO∗
0·66
0·63
0·96
1·07
0·72
1·08
0·51
0·79
0·70
0·75
0·62
0·74
0·23
0·66
0·56
0·44
CaO
12·3
312
·58
11·2
111
·61
12·0
711
·64
11·4
59·
5911
·63
10·0
314
·10
12·9
111
·22
10·4
510
·23
7·36
Na 2
O5·
174·
394·
844·
585·
374·
564·
695·
654·
495·
543·
323·
885·
995·
945·
335·
40
K2O
0·29
0·27
0·50
0·34
0·39
0·40
0·41
0·57
0·39
0·44
0·17
0·21
0·22
0·58
0·31
1·14
Tota
l10
0·07
99·4
398
·95
98·1
310
0·29
99·6
998
·54
98·9
098
·32
98·7
098
·54
98·3
110
0·79
100·
5299
·06
98·5
4
An
56·0
60·3
54·4
57·0
54·1
57·1
55·8
46·5
57·1
48·5
69·2
63·6
50·1
47·5
50·3
39·8
Ab
42·4
38·1
42·5
40·7
43·6
40·4
41·3
49·6
39·9
48·4
29·5
34·7
48·4
48·9
47·5
52·8
Or
1·6
1·5
2·9
2·0
2·1
2·3
2·4
3·3
2·3
2·5
1·0
1·3
1·2
3·1
1·8
7·3
Low
An
143
·452
·954
·343
·754
·147
·647
·245
·344
·733
·469
·255
·850
·144
·339
·937
·8
Hig
hA
n64
·760
·372
·965
·563
·760
·083
·683
·868
·662
·281
·063
·674
·257
·976
·352
·2
1 Low
and
hig
hA
nar
efo
ral
lan
alyz
edp
oin
tsin
ag
iven
sam
ple
.
681
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 8. Ternary components in plagioclase phenocryst cores and rims and groundmass (gmass) plagioclase as a function of host rock composition.
Table 6: Representative amphibole analyses
Sample: MW975 MW974 MW9743 MW9721
core rim core rim core core rim
SiO2 42·44 42·36 43·56 44·19 42·30 42·00 42·73
TiO2 2·62 2·51 2·17 2·11 2·60 2·66 2·18
Al2O3 11·35 11·42 10·36 8·61 12·32 12·47 12·08
Cr2O3 0·17 0·19 0·01 0·04 0·36 0·16 0·12
FeO∗ 12·70 13·10 12·50 12·00 10·70 11·30 12·40
MnO 0·15 0·13 0·18 0·19 0·12 0·10 0·13
MgO 13·57 13·54 14·15 14·35 14·84 14·25 14·16
CaO 10·92 10·99 11·18 11·95 11·32 11·33 11·05
Na2O 2·16 2·21 2·00 1·87 2·39 2·29 2·22
K2O 0·63 0·57 0·60 0·55 0·75 0·76 0·73
F 0·08 0·17 0·35 0·64 0·25 0·25 0·07
Cl 0·05 0·05 0·02 0·05 0·02 0·03 0·01
Total 96·84 97·24 97·08 96·55 97·97 97·60 97·88
O = F;Cl 0·04 0·08 0·15 0·28 0·11 0·11 0·03
Total 96·80 97·16 96·93 96·27 97·86 97·49 97·85
phenocrysts in andesitic rocks range from pargasite to studies demonstrate that AlIV exhibits correlated increaseswith (Na + K)A with rising crystallization temperature.pargasitic hornblende, whereas those in dacitic rocks are
pargasitic hornblende to edenite (Leake, 1978; Haw- The higher AlIV and (Na + K)A in amphiboles from theandesitic rocks relative to most grains in the dacitic rocksthorne, 1981; Table 6).
Figure 7b plots AlIV vs (Na+K) cations in the A site of are consistent with their less evolved nature and inferredhigher liquidus temperatures. A few amphibole crystalsamphibole phenocrysts in Washburn rocks. Experimental
682
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
from the dacitic rocks have cores with high AlIV and (Na of data; (3) at a given SiO2 content, rocks from Mt.+ K)A that plot at the high-temperature end of the Washburn tend to have lower MgO and higher Al2O3,experimental data arrays. These cores probably represent respectively, than rocks from the SW Washburn Range.unequilibrated regions from higher-temperature (or pres- Meen (1985) emphasized plots of K2O vs K2O/MgOsure) stages of evolution because they are compositionally and MgO for discriminating rocks at Independence vol-similar to phenocrysts in the andesitic rocks and are cano (Fig. 2), believed to be related exclusively by high-surrounded by normally zoned rims similar to most other pressure fractional crystallization of mantle-derived par-amphibole crystals in the dacitic lavas. ents in the absence of crustal contamination from those
that interacted with crustal materials. In theory, fractionalcrystallization of magnesian minerals plus plagioclaseleads to production of evolved liquids that fall withinCHEMICAL COMPOSITIONS narrow coronal bands characterized by large increases
We analyzed 46 rocks from Washburn volcano for major in K2O compared with smaller increases in K2O/MgOand trace element abundances by X-ray fluorescence and decreases in MgO (Fig. 10a). In contrast, con-(XRF) spectrometry. Sample locations are shown in Fig. tamination of magmas with low-K2O/MgO crustal melts2. On the basis of the XRF analyses, a set of 17 rocks produces trajectories with low slopes on such plots. Figurewas selected for additional determination of rare earth 10a illustrates a plot of K2O vs K2O/MgO for Washburnelement (REE) and other trace element contents by volcano rocks. Many samples from the SW Washburninstrumental neutron activation analysis (INAA) and a Range define a narrow coronal band nearly identical tofurther subset of these (n = 9) for determination of those formed by suites of rocks at Independence volcanoSr and Nd isotopic ratios by thermal ionization mass inferred by Meen (1985) to be uncontaminated. The SWspectrometry (TIMS). All data are reported in Table 2 Washburn rocks that define this band cannot representand details of the analytical procedures used are described a single uncontaminated liquid line of descent, however.in the Appendix. As discussed below, variations in trace element contents
and isotopic ratios are inconsistent with this proposition.Furthermore, nearly all Washburn rocks, including the
Major elements most mafic samples, form a shallow array in Fig. 10bthat cuts Independence ‘uncontaminated’ trends at highVariations in major element compositions of Washburnangles. We interpret the Washburn array as a mixingrocks are illustrated with respect to SiO2 in Fig. 9.line between high-MgO basaltic magmas and low-MgOAccording to the classification schemes of Peacock (1931),crustal melts. This implies that the compositions of allPeccerillo & Taylor (1976), and Le Maitre (1989), Wash-Washburn rocks were affected by crustal interactionburn igneous rocks form a medium- to high-K, calc-
alkaline suite (alkali–lime index>60; not shown) ranging and none can be assumed to be direct uncontaminatedin composition from basaltic andesitic through dacitic. differentiates of primary mantle-derived magmas.In these respects the compositions of Washburn rocksshare broad overall similarities with other Eocene calc-alkaline igneous rocks in the Absaroka Volcanic Province
Trace elementsand elsewhere in the ‘Challis arc’ (e.g. Norman & Mertz-Variations in trace element abundances of Washburnman, 1991; Hooper et al., 1995; Dostal et al., 1998;volcanic rocks display an interesting set of relationshipsLindsay & Feeley, 1999; Morris et al., 2000). As is typicalthat are similar to, but somewhat more complex thanof calc-alkaline suites, MgO, Fe2O3, CaO, TiO2, andfor major elements. The important features are as follows.MnO decrease with increasing SiO2, whereas Na2O and
Many trace elements (e.g. Ba, Rb, Y, Zr, Sc) displayK2O increase. Trends for Al2O3 and P2O5 are diffuseoverall linear, but diffuse variations with respect to SiO2,and show no obvious correlation with SiO2.although highly compatible trace elements such as CrWith respect to identification of magmatic sources,and Ni scatter widely so that some intermediate com-differentiation processes, and compositional evolution ofposition rocks have relatively high concentrations ofthe Washburn magmatic system, the important majorthese elements (Fig. 11). These features do not reflectelement features are: (1) samples from the SW Washburnxenocrystic olivine as there is no correlation betweenRange consist mainly of a bimodal package of basalticmodal percent olivine and Ni and Cr contents in theandesitic and dacitic rocks, whereas stratigraphicallyandesitic and dacitic rocks (Table 2). They are also nothigher units on Mt. Washburn range continuously fromdue to contamination during sample preparation becausebasaltic andesitic through dacitic rocks, but are dom-the samples were crushed in tungsten-carbide grindinginated by andesitic compositions; (2) trends with respectbowls, which does not cause analytically significant con-to SiO2 for many elements are linear, except for MgO,
Al2O3, and P2O5, which have more diffuse distributions tamination for Ni and Cr (or MgO; Johnson et al., 1999).
683
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 9. Major element compositions of Washburn igneous rocks vs SiO2. K2O classification boundaries are from Peccerillo & Taylor (1976).Total alkali diagram shows classification scheme of Le Maitre (1989). Open symbols are for samples from the SW Washburn Range and filledsymbols are for samples from Mt. Washburn. Circles, squares, and triangles are for basaltic andesitic, andesitic, and dacitic composition rocks,respectively. Dashed arrows through dacitic composition rocks on MgO vs SiO2 diagram are linear regression lines.
It is noteworthy that, analogous to variations in MgO, zircon (or crustal melting with these phases residual) werenot major processes in the evolution of the Washburnrocks from Mt. Washburn generally have lower Ni and
Cr contents than most equivalent bulk composition rocks suite as a whole. In contrast, Y decreases with increasingSiO2, indicating that fractionation of large amounts offrom the SW Washburn Range. Moreover, the samples
that form a narrow coronal band in Fig. 10a do not clinopyroxene or amphibole, or contamination by crustalmelts with low Y contents was important. On the basisdefine a single liquid line of descent resulting from
fractionation of magnesian minerals plus plagioclase in of location and Sr, Zr, Y, and Rb contents, three groupsof dacitic rocks can be distinguished at Washburn vol-diagrams illustrating variation of Ni and Cr with SiO2.
Accordingly, diagrams such as Fig. 10a may not uniquely cano; two from the SW Washburn Range, designatedtypes 1 and 2, and one from Mt. Washburn, designatedidentify magmas strictly related by simple closed-system
fractional crystallization (see Meen, 1985; Meen & Eggler, type 3. On a comparative basis, type 1 dacites (includingone high-silica andesitic lava) have higher Rb, Zr, and1987).
Both Sr and Zr increase slightly with increasing SiO2 Y, and lower Sr contents than type 2 dacites (Fig 10).Type 3 dacites have similarities to and differences fromcontents, indicating that fractionation of plagioclase and
684
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
the decreasing HREE contents with increasing SiO2 wt %virtually require that differentiation at Washburn volcanoinvolved mixing with, or assimilation of, crustal melts withlow HREE contents. They cannot reflect fractionationof substantial amounts of amphibole, the only majorphenocryst phase in Washburn rocks with mineral–meltKd values >1 for the MREE and HREE. Fractionationof substantial amphibole results in depletion of MREErelative to HREE and the production of concave-upwardREE patterns.
Mid-ocean ridge basalt (MORB)-normalized trace ele-ment plots for Washburn basaltic andesitic rocks arecharacterized to some extent by features considered diag-nostic of subduction-related magmas (Fig. 13; Pearce,1982). For example, relative to MORB, basaltic andesiteshave strong enrichments in large ion lithophile elements(LILE; Sr, K, Rb, Ba, and Th) and depletions in thehigh field strength elements (HFSE) Ta, Nb, and Ti. TheHFSE Zr and Hf, however, are not depleted relative toMORB. High-K2O basaltic rocks from the shoshoniticSunlight volcano, eastern AVP (Fig. 1), are further en-riched in LILE, but not HSFE, relative to Washburnbasaltic andesites (Fig. 13b). Least evolved (MgO
Fig. 10. (a) K2O vs K2O/MgO variation diagram for Washburn rocks. 5·93 wt %; Meen & Eggler, 1987) high-alumina tholeiiticThe two continuous curves enclose Washburn data defining a narrow basaltic andesite from Independence volcano, easterncoronal band similar to bands formed by Independence volcano samples
AVP, shows variable enrichments and depletions in LILEbelieved to be free from crustal contamination (Meen, 1985). As anrelative to Washburn basaltic andesites, although with theexample, the two dashed curves enclose ‘uncontaminated’ samples from
the Independence stock (data not shown). Arrows show inferred effects exception of Ti, HFSE are depleted. Plots for Washburnof fractional crystallization and crustal contamination. (b) K2O vs MgO andesitic and dacitic rocks are enriched in both LILEvariation diagram for Washburn rocks. Area enclosed by dotted line
and Zr and Hf relative to basaltic andesitic rocks, althoughis the field for Independence volcano from Meen (1985). Arrows showtrends for Independence samples believed to be free from crustal they are depleted in Ti, Y, and Yb.contamination (see Meen, 1985). Data symbols in both panels as inFig. 9. (See text for further discussion.)
Sr and Nd isotopesInitial 87Sr/86Sr and 143Nd/144Nd isotopic ratios correctedtypes 1 and 2 in that they have Y contents similar tofor 53 my of in situ growth of radiogenic Sr and Nd forthose of type 2 dacites, Sr contents similar to or greaternine Washburn rocks are presented in Table 2 andthan those of type 1 dacites, and Zr and Rb contentsillustrated in Fig. 14. Relative to ‘bulk Earth’ all rocksintermediate between those of both types.have high Sr and low Nd isotopic ratios and thus plotAll Washburn volcanic rocks have LREE-enriched,within the ‘enriched’ quadrant in Fig. 14a. The datachondrite-normalized REE patterns (Fig. 12). For thedefine a negatively correlated array extending from asuite as a whole andesitic and dacitic rocks tend to havefield defined by ultramafic xenoliths brought up in Eocenelower concentrations of HREE, higher concentrations ofalkalic magmas in the Crazy Mountains,>150 km northLREE, and flatter middle (MREE) to HREE patternsof Washburn volcano, to fields defined by Wyoming(e.g. Tb–Lu) than basaltic andesitic rocks. As a result ofProvince Archean granulite-facies rocks brought up asthe first two features chondrite-normalized REE plotsxenoliths in Cenozoic magmas (Fig. 14a; Leeman et al.,display fanning (cross-over) patterns with fulcrums1985; Dudas et al., 1987) and Archean supracrustalcentered between Nd and Sm. All Washburn rocks haveamphibolites and granitoids that dominate the exposedeither no Eu anomaly or small positive anomalies thatbasement near Washburn volcano (Meen, 1987a;increase from 1·00–1·03 for basaltic andesitic rocks toWooden & Mueller, 1988). Washburn mafic rocks have1·01–1·15 for andesitic and dacitic rocks. These re-slightly higher 87Sr/86Sr ratios but similar 143Nd/144Ndlationships are consistent with the increasing Sr con-ratios compared with mafic rocks from Independencecentrations with SiO2, in that both indicate thatand Sunlight volcanoes in the eastern belt of the AVPfractionation of plagioclase was not a major process
during evolution of the Washburn suite. Furthermore, (Meen & Eggler, 1987; T. C. Feeley, unpublished data,
685
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 11. Selected trace element contents in Washburn igneous rocks vs SiO2. Symbols as in Fig. 9. Symbols with center point in Ni and Crdiagrams are those defining coronal band in Fig. 10a. Dacitic rock types are distinguished on the basis of location and Sr, Zr, Y, and Rbcontents. Dashed arrows through dacitic composition rocks in Ni and Cr vs SiO2 diagrams are linear regression lines.
2001). Although the Washburn rocks have MORB-nor- isotopic ratio than the single analyzed type 2 dacitic rock.This feature is consistent with the relatively high Rb/Srmalized trace element patterns characterized by de-
pletions in Ta and Nb (Fig. 13), they plot well below ratios for type 1 dacitic rocks as a whole (e.g. Fig. 11).Together these relationships imply that the continentalthe field defined by modern subduction-related basalts
because of their low Nd isotopic ratios. This feature is crust beneath Washburn volcano is isotopically hetero-geneous but remained isolated for significant periods ofcharacteristic of many early- to late-Cenozoic magmatic
rocks in the Wyoming Province and is considered to time to allow for greater in situ growth of radiogenic Srin Rb-enriched sources for type 1 dacitic rocks. It isreflect derivation from or interaction with ancient sub-
continental lithospheric mantle (e.g. Fraser et al., 1985; also noteworthy that Washburn basaltic andesitic andandesitic rocks contained within the coronal band in Fig.Dudas et al., 1987; Meen & Eggler, 1987; Scambos, 1991;
MacDonald et al., 1992; O’Brien et al., 1995). 10a have variable Nd and Sr isotopic compositions thatcorrelate with SiO2. Such relationships are typical ofThe Nd and Sr isotopic data indicate that crustal
material was important in the genesis of Washburn many other volcanic suites resulting from assimilation ofor mixing with continental crust and they preclude anmagmas, and differences in crustal composition may
provide an explanation for the contrasts between the types origin for Washburn basaltic andesites by closed-systemprocesses. Accordingly, although diagrams such as Fig.of dacitic magmas described previously. For example,
although values for all rock types overlap, there are fairly 10a may be useful for identifying magmas having ex-perienced some fractional crystallization during theirwell defined positively and negatively correlated arrays
between SiO2 content and Sr and Nd isotopic com- genesis, they do not uniquely preclude the possibility thatthese magmas have also experienced contamination bypositions, respectively (Fig. 14b and c). In addition, at
least one type 1 dacitic rock has a significantly higher Sr continental crust (see Meen, 1985; Meen & Eggler, 1987).
686
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
data for these lithologies are therefore not rigorouslyconstrained because trace element and isotopic cal-culations are based on data from different sources andlocalities. Specifically, the trace element calculations util-ize data from Joswiak (1992), whereas the isotopic cal-culations utilize data from Leeman et al. (1985) becausethese papers contain the most comprehensive trace ele-ment and isotopic datasets, respectively, for deep-crustallithologies in the Wyoming Province. The models do,however, place constraints on processes and sourcesresponsible for the origin of compositional diversity inthe Washburn magmatic system.
The second limitation is that Washburn mafic rockshave mg-numbers (= 100[Mg/(Mg + Fe∗)]) <64 andNi <175 ppm. Thus, none represent primary mantlemelts. Furthermore, primary mantle melts are generallyunknown from the AVP. Meen & Eggler (1987) andMeen (1985) argued on the basis of plots of K2O vsK2O/MgO (e.g. Fig. 10a) that rocks from the centralportion of the Independence stock represent un-contaminated mantle-derived melts related solely by frac-tional crystallization. In support of this conclusion theypresented Sr (0·70453–70475) and Nd (0·51182–0·51196)isotopic data for these rocks virtually constant acrossthe compositional spectrum (basaltic andesitic to high-K
Fig. 12. Rare earth element abundances of Washburn rocks nor- dacitic). This conclusion is now unclear, however, becausemalized to nonvolatile C1 chondrite from Anders & Ebihara (1982).
Meen (1985) and Meen & Eggler (1987) calculated initialGd∗ is extrapolated from the LREE trend. Symbols as in Fig. 9.isotopic ratios for the Independence rocks using an ageof 84 Ma. Recently reported 40Ar/39Ar data on mineralseparates clearly show that the age of the IndependencePETROGENESIS OF WASHBURNstock is 49·96–48·50 Ma (Harlan et al., 1996). Using these
VOLCANO MAGMAS ages the calculated initial Sr (0·70459–0·70508) and NdIn the following sections several models for the origin (0·51167–0·51192) isotopic data for the Independenceof compositional diversity in the Washburn system are stock show analytically significant variations and modestevaluated. In these models we use the geochemical and correlations exist with parameters such as K2O, SiO2,isotopic data in Table 2, published data available for and 1/Sr.potential crustal source components, and standard ex- To address the lack of primary melts erupted at Wash-pressions for (1) trace element and isotopic evolution burn volcano, as a first step in the modeling we calculatedduring Rayleigh fractionation of crystals from magma, a hypothetical primary mantle melt composition to use(2) simultaneous fractional crystallization and wall as- as a starting point. The composition of the primary meltsimilation (DePaolo, 1981), and (3) incremental batch was estimated by assuming that (1) it has 400 ppm Nimelting of crustal rocks. As discussed below, there are (Sato, 1977; Frey et al., 1978; Wilson, 1989) and (2) atwo significant limitations on the models. high Ni, but olivine-free (to avoid cumulate effects) andes-
The first limitation is that there is no direct evidence itic rock (MW9729) is a simple two-component mixturefor the composition of crustal rocks beneath Washburn of an evolved dacitic magma (OP9881) and the primaryvolcano. Published geochemical data are available for melt. Using the proportions of the endmembers predictedrocks from nearby areas, including information on the by this procedure (68:32) yields a basalt with 49·3 wt %compositions of supracrustal igneous and metamorphic SiO2, 14·6 wt % MgO, and an mg-number of 68, whichrocks in the Beartooth Mountains and Wyoming Province are not unreasonable estimates for primary or primitivedeep-crustal granulite-facies xenoliths (Mueller et al., calc-alkaline magmas that were in equilibrium with1982, 1983, 1985; Wooden et al., 1982; Leeman et al., mantle peridotite (Gill, 1981; Foden, 1983; DeBari &1985; Dudas et al., 1987; Meen, 1987a; Wooden & Sleep, 1991).Mueller, 1988; Collerson et al., 1989; Joswiak, 1992). In Fig. 13 the MORB-normalized trace element patternOnly small, partial datasets exist for compositions of the of the model primary melt is compared with plots for
several other mafic rocks in the AVP and Challis–granulite-facies xenoliths, however. The models utilizing
687
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 13. MORB-normalized trace element diagrams. (a) Trace element abundance patterns for rocks from Mt. Washburn. (b) Trace elementabundance patterns for rocks from the SW Washburn Range contrasted with mafic basaltic andesite from Independence volcano (2080; Meen& Eggler, 1987) and high-K basalt from Sunlight volcano (T. C. Feeley, unpublished data, 2001). (c) The range in trace element abundancesfor model Washburn primary magma (method of calculation in text) contrasted with subalkaline mafic rocks from the Colville Igneous Complex(Morris et al., 2000), the Challis volcanic field (Norman & Mertzman, 1991), and the Crazy Mountains (Dudas, 1991). (d) Model Washburnprimary magma contrasted with high-K basalt from Sunlight volcano (T. C. Feeley, unpublished data, 2001), basaltic andesite from Independencevolcano (Meen & Eggler, 1987), and alkaline basalt from the Crazy Mountains (Dudas, 1991). In (a) and (b), symbols as in Fig. 9. It should benoted in (c) and (d) that not all elements are available for all samples. In all diagrams normalizing values and element order after Pearce (1982).Also shown for reference in all diagrams are typical patterns for tholeiitic ocean-island basalt (OIB) and calc-alkaline volcanic arc basalt (Pearce,1982).
Absaroka volcanic episode (e.g. Figs 1 and 2). These basalt has nearly identical Nb, Zr, Ti, and Y contents tothe model Washburn basalt, but LILE are greater by asinclude subalkaline basalts from the Challis volcanicmuch as a factor of three. The Independence basalticfield (Norman & Mertzman, 1991), the Colville igneousandesite has lower concentrations of K, Rb, and Th, butcomplex (Morris et al., 2000), and the Crazy Mountains,concentrations of Sr, Ba, and P that exceed those in theMontana (Dudas, 1991; Fig. 13c); a high-alumina tholei-Washburn basalt by a factor of two. The alkaline basaltitic basaltic andesite from Independence volcano (Meenfrom the Crazy Mountains is enriched in nearly all trace& Eggler, 1987); a potassic trachy-basalt from Sunlightelements compared with the subalkalic rocks.volcano, eastern AVP (T. C. Feeley, unpublished data,
2001); and an alkaline basalt (malignite) from the CrazyMountains (Fig. 13d).
Origin of intermediate compositionIn general, the trace element pattern for the modelmagmasWashburn basalt is similar to patterns for the subalkaline
basalts from the Colville igneous complex, the Challis To illustrate processes responsible for producing thevolcanic field, and Crazy Mountains, suggesting melting spectrum of intermediate composition magmas at Wash-of similar sources (Fig. 13c). Only minor differences in burn volcano, Fig. 15 presents several calculated differ-most elemental concentrations are evident, the major entiation models as a function of trace element variationsexception being Nb, for which the other basalts have vs Sc. For these models Sc was selected as a trace elementsignificantly higher concentrations. Patterns for the other index of differentiation because there is a well-correlatedmafic rocks show more pronounced differences from the decrease with increase in SiO2, and Sc emphasizes theWashburn model primary basalt pattern and may derive role of clinopyroxene, the most abundant mafic silicate
phase in the rocks.from more disparate sources (Fig. 13d). The Sunlight
688
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Fig. 14. (a) Comparison of initial Sr and Nd isotopic compositions (at 53 Ma) of Washburn igneous rocks with published data for crustal rocksfrom the Wyoming Province and Eocene igneous rocks from the AVP. Highwood Mountains xenolith field from Joswiak (1992), Crazy Mountainsxenolith fields from Dudas et al. (1987), Snake River Plain xenolith field from Leeman et al. (1985), Beartooth Mountains Archean granitoid andamphibolite field from Meen (1987a) and Wooden & Mueller (1988), Independence fields from Meen & Eggler (1987), and Sunlight volcanofield from T. C. Feeley (unpublished data, 2001). All data are calculated for an age of 53 Ma except for data from Independence and Sunlightcenters. Data defining these fields calculated for 50 Ma for the Independence volcano (Harlan et al., 1996) and 48 Ma for the Sunlight volcano(Feeley et al., 1999). (b) ( 87Sr/86Sr)i and (c) ( 143Nd/144Nd)i vs SiO2 for Washburn rocks. Symbols as in Fig. 9. Symbols with center point in (b) and(c) are those defining coronal band in Fig. 10a.
replicate the range of compositions observed. This processThe continuous curves labeled ‘fx’ originating from theis successful in accounting for the contrasting wedge-model primary basaltic magma in Fig. 15 are calculatedshaped data fields in Fig. 15 defined by large variationsfractional crystallization trends assuming a mineral as-in Ni and Cr contents for basaltic andesitic rocks on thesemblage of 40% clinopyroxene, 30% olivine, 29%one hand, and large variations in LILE and Y for daciticplagioclase, and 1% Fe–Ti oxides, and intermediaterocks on the other. Figure 16 tests the validity of thevalues for the range of basalt mineral–melt partitionmixing model for the andesitic magmas using ratio–ratiocoefficients presented by Rollinson (1993). This plagio-variation diagrams and corresponding companion plotsclase-poor, augite-rich phenocryst assemblage is char-following the method of Langmuir et al. (1978). Excludingacteristic of crystallization under conditions of elevatedtype 1 dacitic rocks, the diffuse, but hyperbolic form ofpressure following the experimental work of Sisson &the data array is consistent with a mixing relationshipGrove (1993). We do not illustrate nor do we considerbetween diverse composition mafic and silicic magmasfractionation of modal proportions of phenocrysts as ato produce the andesitic magmas (Langmuir et al., 1978).suitable process because all rocks contain low-pressure,For example, in Fig. 16a four calculated possible mixingplagioclase-rich mineral assemblages that if removedarrays are illustrated. Curves 1a and 1b illustrate mixingwould increase Sc and deplete Sr contents in residualbetween the most primitive sample in the Washburnliquids, opposite to the depletion in Sc and increase insuite (MW9713) with average compositions of type 2 andSr contents with increasing SiO2 observed for the suite.type 3 dacites (diamonds). Curves 2a and 2b illustrateIn addition, the increasingly positive Eu anomalies andmixing with a slightly more evolved basaltic andesitedecreasing HREE with increasing SiO2 also argue(MW9716). The close correspondence of the data to thestrongly against any process involving shallow pressurecalculated curves supports the mixing model and impliesfractionation.that type 1 dacitic magmas were not utilized as silicicThe model curves in Fig. 15 indicate that small amountsendmembers. Figure 16b and c are companion plots inof fractionation (e.g. F = amount original magma re-which the ratios in Fig. 16a are plotted against a ratiomaining >0·8) followed by mixing with diverse com-
position dacitic magmas (thick, short-dashed lines) can of the denominators (see Langmuir et al., 1978). The
689
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Fig. 15. Select trace element contents for Washburn rocks vs Sc contents. Data symbols as in Fig. 9. The large open circle is the compositionof a calculated hypothetical primary basaltic magma (for method of calculation see text). The continuous curves labeled fx originating from theprimary basaltic magma are calculated fractional crystallization trends assuming intermediate values for the range of mineral–melt partitioncoefficients presented by Rollinson (1993). Tick marks next to fractional crystallization and assimilation plus fractional crystallization (afc) trendsrepresent 10% intervals of fractionation (= fraction of original magma remaining). (See text for further discussion.)
consistency of the mixing relationship is substantiated by value is considered to simulate differentiation under mid-to deep-crustal conditions where the crust is at an elevatedthe linear variation of the data (excluding type 1 dacites)
in these diagrams (Langmuir et al., 1978). temperature and olivine, clinopyroxene, and plagioclaseare multiply saturated (Reiners et al., 1995). Additionally,It is also likely that assimilation of crustal rocks occurred
during fractional crystallization of the mafic parental we included 10% garnet in the fractionating assemblagefor the deep-crustal assimilation model, to investigatemagmas, as these processes are inferred to be physically
coupled (Taylor, 1980) and probably accounts for the assimilation at high pressures where garnet, with elevatedpartition coefficients for Sc (2·6–3·7; Irving & Frey, 1978;absence of data points in direct proximity to the model
closed-system fractionation curves in Fig. 15 and the Hauri et al., 1994) and Cr (nine; Johnson, 1998), is animportant residual phase during crustal melting.isotopic diversity among the basaltic andesitic rocks. As
a first-order approximation of the effects of assimilation The model assimilation–fractional crystallizationcurves suggest, like the closed-system trends, that onlyplus fractional crystallization, we constructed model
curves using the method of DePaolo (1981), the crys- small amounts (F >0·9) of differentiation of a potentiallyprimary magma are necessary to produce the range oftallizing assemblage discussed above, and crustal as-
similants comparable in composition with the average Ni and Cr contents in parental mafic liquids required inthe subsequent two-component mixing models. The mostof supracrustal Archean granitoids (curve afc 1) and
amphibolites (curve afc 2) in the Beartooth Mountains important differences between the supracrustal (curvesafc 1 and afc 2) and deep-crustal (curve afc 3) assimilation[data of Meen (1987a)] and deep-crustal granulite-facies
xenoliths in the Wyoming Province (curve afc 3). The r models is that LILE such as Rb and Ba are elevatedmore rapidly in the former whereas Y contents arevalue (ratio of rates of assimilation to crystallization; see
DePaolo, 1981) assumed during all models is 0·8; this depleted in the latter. However, none of the models
690
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Fig. 16. (a) Sc/Zr vs Sr/Y variation diagram for Washburn rocks following the method of Langmuir et al. (1978). Symbols as in Fig. 9. Curves1a and 1b illustrate mixing between the most primitive sample in the Washburn suite (MW97-13) with average compositions of type 2 and type3 dacites (white and gray diamonds, respectively). Curves 2a and 2b illustrate mixing with a slightly more evolved basaltic andesite (MW97-16).(b) and (c) are companion plots in which each of the ratios in (a) is plotted against a ratio of the denominators. Correlation coefficients = r(excluding type 1 dacitic rocks). It should be noted that type 1 dacitic magmas do not appear to have been utilized as silicic endmembers in themixing process.
illustrated in Fig. 15 can independently reproduce theunique compositional features of the Washburn suite.More complex mathematical expressions for the as-similation–fractional crystallization process have alsobeen considered (Spera & Bohrson, 2001). These alsoare unable to independently reproduce the roughly lineartrends for many elements and the high Cr and Ni contentsof the andesitic and dacitic rocks. Together these featuresvirtually require mixing between mafic and silicic magmasto produce intermediate composition magmas (McMillan& Dungan, 1984, 1986; Bacon, 1986).
The relationships illustrated in Fig. 15 do not uniquely Fig. 17. Calculated assimilation–fractional crystallization and mixingmodels for ( 87Sr/86Sr)i and ( 143Nd/144Nd)i using various crustal end-constrain the crustal lithologies assimilated because ofmember compositions. Symbols same as in Fig. 9. Tick marks next tothe limited amounts of differentiation of the Washburncurves represent fraction of original magma remaining. (See text forparental magmas. Therefore, to better constrain the further discussion.)
identity of crustal lithologies potentially assimilated duringdifferentiation of Washburn parental magmas, as-
1987a) indicate elevated Sr isotopic ratios and bulk Srsimilation plus fractional crystallization calculations werecontents, which produce strong enrichments in 87Sr/carried out with Sr and Nd isotopic data. The results of86Sr relative to 143Nd/144Nd during the early stages ofa variety of calculations are shown in Fig. 17, whichdifferentiation. Instead, the large range in Nd isotopicrepresent the best of a family of models in which com-compositions at roughly constant 87Sr/86Sr ratios ofpositions of the assimilants were chosen to be consistentWashburn basaltic andesitic magmas requires as-with known crustal lithologies. These calculations pro-similation of crustal rocks with low 87Sr/86Sr and 143Nd/duce realistic constraints on possible crustal assimilants144Nd ratios. Rocks with these features occur as granulitebecause Nd is highly incompatible during differentiationxenoliths brought to the surface in late Cenozoic basalticof mafic magmas and the Sr bulk distribution is controlledmagmas of the Snake River Plain in the Spencer–Kilgoreprimarily by plagioclase.area >125 km to the SW of Washburn volcano (afc 3;Assimilation plus fractional crystallization calculationsLeeman et al., 1985). Assimilation of rocks with theseusing the average of supracrustal Archean granitoidsisotopic features results in little change in Sr isotopic(curve afc 1; Fig. 17) and amphibolites (curve afc 2) inratios, although Nd isotopic ratios are lowered drasticallythe Beartooth Mountains cannot produce the 87Sr/86Sr(afc 3; Fig. 17). This process produces a range of com-vs 143Nd/144Nd trend for Washburn basaltic andesitic
rocks. Published data for these crustal rocks (Meen, positions that are appropriate for mixing with Washburn
691
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
dacitic magmas to produce the intermediate composition the majority of rocks from the SW Washburn Rangesuite with anomalously high MgO, Ni, and Cr contentsmagmas. Therefore, a model where Washburn parental
magmas undergo combined assimilation–fractional crys- on the one hand, and rocks from Mt. Washburn withlower contents of these elements on the other (Fig. 15).tallization in the deep crust followed by mixing with
silicic melts is consistent with the model calculationsperformed here.
Crustal melting, restite unmixing, magma mixing, andfractional crystallization
Geochemical characteristics (as discussed above) of theOrigin of Washburn dacitic magmas Washburn dacitic magmas are hard to reconcile by a
fractionation- or assimilation-dominated process, addingThe conclusion that Washburn intermediate compositioncredence to a crustal melting hypothesis for their origin.rocks are products of mixing between mafic and silicicAssuming this to be correct, the origin of the roughlymagmas does not preclude an origin for the daciticlinear trends for individual groups of dacitic rocks inmagmas by a fractionation-dominated liquid line of des-some Harker plots (e.g. Al2O3; Fig. 9) may reflect severalcent from less evolved magmas, as several workers haveprocesses, including crystal fractionation, restite un-proposed mechanisms capable of producing frac-mixing, mixing with less evolved magmas, and meltingtionation-generated composition gaps in calc-alkaline sys-processes occurring within source regions.tems (e.g. Grove & Donnelly-Nolan, 1986; Brophy, 1991).
Mixing between crustally derived silicic melts andVarious geochemical criteria indicate that the daciticmagmas in the Washburn suite (and hence, the silicic restite assemblages as a petrogenetic process is typically
reserved for coarse-grained, felsic plutonic rocks (e.g.mixing endmembers) are, however, not direct frac-tionation products of basaltic andesitic magmas at present White & Chappell, 1977). For rocks in the Absaroka
Volcanic Province this process was discussed by Meenexposed at the volcano. The strongest and most obviousevidence against a closed-system process is that dacitic & Eggler (1989) for granitoids associated with the In-
dependence volcanic center. In their study of the In-rocks have lower Nd and higher Sr isotopic ratios thanthose of most less evolved rocks. Thus, if fractional dependence granitoids, Meen & Eggler (1989) argued
that phenocryst assemblages represent restite retainedcrystallization produced the dacitic magmas, it was ac-companied by crustal assimilation. from zones of crustal melting and silicic melt production.
Although broadly similar in composition to the In-Geochemical characteristics of Washburn daciticmagmas are also difficult to reconcile with an origin by dependence granitoids, chemical variations of the Wash-
burn dacites probably do not reflect this process, becauseassimilation of crustal rocks plus fractional crystallizationof mafic magmas. The strongest evidence against an of the lack of petrographic evidence that the phenocryst
assemblages represent non-magmatic, residual as-assimilation process is the high contents of compatibletrace elements (e.g. Cr 159–42 ppm; Ni 63–25 ppm) and semblages, such as highly corroded mineral phases and
high-pressure, high-temperature, anhydrous mineralsthe large ranges in incompatible trace element contentsin dacitic rocks from Washburn volcano. Together these (Clemens, 1989). Furthermore, dacitic magmas at Wash-
burn volcano with similar major element compositionsfeatures conflict with an origin by Rayleigh-type frac-tionation coupled with assimilation as this process gen- have wide ranges in modal phenocryst proportions, in-
consistent with a restite unmixing process (e.g. Fig. 5).erates coupled exponential enrichments and depletions,respectively, in these trace element groups. It is possible Restite unmixing is also unlikely on the basis of chem-
ical variations. For illustration, following a similar (butthat low contents of compatible trace elements in somedacitic rocks from Mt. Washburn in part reflect extensive less rigorous) procedure to that employed by Meen &
Eggler (1989), we estimated SiO2 contents of modelfractionation. However, model assimilation curves in Fig.15 predict irreconcilable incompatible and compatible crustal melts by extrapolating linear regression lines for
MgO, Ni, and Cr contents of the SW Washburn Rangetrace element contents using any reasonable crustal con-taminant. As described above, the simplest interpretation and Mt. Washburn dacites to zero intercepts on Harker
diagrams (Figs 9 and 10). Using the average SiO2 contentof the low contents of compatible trace elements isthat they reflect mixing with mafic endmember magmas for each group as predicted by the intercepts, contents
of other major elements in model crustal melts wereslightly more fractionated than those for the SW Wash-burn Range suite. Thus, all constraints together indicate estimated by extrapolation of regression lines (not shown)
to these values. Table 7 presents the model crustalthat the net result of fractionation plus assimilation atWashburn volcano is not to produce highly evolved melt compositions and the results of least-squares mixing
calculations involving the melts and phenocryst phasesmagmas, but to contribute to the genesis of a spectrumof mafic magma types and the formation of multiple in the Washburn dacitic rocks. The results indicate that
dacitic magmas broadly similar in composition (�r2 =mixing lines. The main mixing trends are identified by
692
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Table 7: Major element compositions of model crustal melts and calculated residual mineral assemblages for
dacitic lavas
Model (1) SW Washburn Range (2) Mt Washburn
Crustal Rock2 Calc. Residual Solution wt % Crustal Rock Calc. Residual Solution wt %
melt1 MW975 rock3 phases4 melt MW9739 rock phases
wt %5
SiO2 71·68 65·27 64·59 0·68 Plag 11·0 68·38 63·39 62·68 0·71 Plag 10·0
TiO2 0·31 0·51 0·70 –0·19 Opx 2·2 0·31 0·61 0·71 –0·10 Amph 13·9
Al2O3 15·61 16·34 16·02 0·32 Amph 15·1 17·37 17·27 17·49 –0·22 Mag 0·9
Fe2O3 2·06 4·60 4·53 0·07 Mag 0·6 2·84 5·14 5·06 0·08 F 75·2
MnO 0·03 0·07 0·05 0·02 F 6 71·1 0·03 0·05 0·05 0·00 �r2 1·27
MgO 0·00 2·66 2·75 –0·09 �r2 1·51 0·00 1·80 2·03 –0·23
CaO 1·96 4·46 4·33 0·13 2·85 5·17 4·80 0·37
Na2O 5·14 3·72 4·64 –0·92 5·18 4·07 4·78 –0·71
K2O 3·08 2·17 2·31 –0·14 2·85 2·29 2·25 0·04
P2O5 0·14 0·20 0·10 0·10 0·20 0·21 0·15 0·06
1Method of calculation described in text.2Composition of mafic dacite modeled as crustal melt + restite. For SW Washburn Range, MW975 is used because ofcorresponding microprobe analyses of mineral phases (Tables A4–A6). For Mt. Washburn, composition is that of most maficdacite.3Best fit calculated major element composition of dacite.4Model restite assemblage.5All compositions normalized to 100% anhydrous.6Percentage of crustal melt in rock calculated by model.�r 2 is sum of the square of the residuals.
1·3–1·5) to the Washburn dacitic rocks can be produced as a whole. This indicates that if fractional crystallizationoccurred, it followed mixing of mafic magmas with silicicby phenocryst addition. However, the models require
similar quantities of amphibole (14–15 wt %) for suites crustal melts compositionally similar to dacitic rocks fromMt. Washburn.of dacitic rocks from the SW Washburn Range and Mt
Washburn, which cannot be reconciled with the differenttrends in Ni and Cr in Figs 11 and 14 for these two
Constraints on crustal source materialssuites and the large variation in modal amphibole in theSr and Nd isotopic ratios of most Washburn andesiticdacitic rocks.and dacitic rocks project toward or plot within fieldsRather than restite unmixing, some of the geochemicalformed by Wyoming Province deep-crustal granulite-variations for the Washburn dacitic rocks may reflectfacies rocks brought up as xenoliths in Cenozoic magmaslimited fractional crystallization of intermediate to silicic(Fig. 14a; Leeman et al., 1985; Dudas et al., 1987). Acomposition magmas following magma mixing. In par-straightforward interpretation of the isotope data is thatticular, crystal fractionation may have been involved inthe sources for the silicic magmas are similar in com-the petrogenesis of SW Washburn dacitic rocks becauseposition to the ancient deep-crustal rocks. For type 1they have Al2O3 (and P2O5) contents that decrease slightlydacitic rocks, supracrustal granitoids and amphibolites inwith increasing SiO2 (Fig. 9). Least-squares mass balancethe Beartooth Mountains might also provide suitablecalculations allow such a process in that they predict
small sums of the squared residuals for major elements crustal sources, as the isotopic composition of the singleanalyzed sample lies at the edge of the field formed byand plagioclase-dominated, low-pressure phenocryst as-
semblages similar to those found in the SW Washburn the compiled data of Meen (1987a) and Wooden &Mueller (1988) in Fig. 14.dacitic rocks (Table 8). It is also notable that the trend
for these rocks in a plot of Al2O3 vs MgO (Fig. 18) Trace element compositions of the dacitic rocks provideadditional evidence that source rocks of the daciticextends off the main mixing trend for the Washburn suite
693
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Table 8: Major element fractional crystallization
model for SW Washburn dacitic lavas
Parent Daughter Calc. Residual Solution wt %
MW97111 MW9742 daughter phases3
wt %4
SiO2 64·63 67·94 68·18 –0·24 Plag 16·5
TiO2 0·52 0·44 0·51 –0·07 Opx 6·3
Al2O3 16·24 15·50 15·37 0·13 Amph 2·8
Fe2O3 4·87 4·36 4·37 –0·01 Mag 0·5
MnO 0·08 0·05 0·08 –0·03 Apatite 0·2
MgO 3·05 1·63 1·61 0·02 F 5 73·7
CaO 4·65 3·49 3·53 0·04 �r2 0·18
Na2O 3·88 3·79 3·86 –0·07
K2O 1·90 2·64 2·34 0·30
P2O5 0·19 0·15 0·15 0·00
1Least silicic SW Washburn dacitic rock.2Most silicic SW Washburn dacitic rock.3Compositions of silicate mineral phases in Tables A4–A6.
Fig. 19. (a) Th vs Zr and (b) Th vs Zr/Th for Washburn volcano4All compositions normalized to 100% anhydrous.dacitic rocks and major crustal lithologies in the surrounding region.5Fraction of liquid remaining.Symbols as in Fig. 9. Data for deep-crustal granulite-facies xenoliths�r2 is sum of the square of the residuals.from Joswiak (1992). Data for Archean amphibolites and granitoidsfrom Mueller et al. (1982, 1983, 1985), Wooden et al. (1982) and Meen(1987a). Fractionation vectors for mineral phases that may control theseelements are shown in (b). cpx, clinopyroxene; amph, amphibole; plag,plagioclase. Trend labeled PM is calculated partial melting trend forcrustal source with 75 ppm Zr and 1·25 ppm Th. Tick marks on curverepresent percent partial melting. (See text for further discussion.)
(e.g. Zr 75 ppm; Th 1·25 ppm). In calculating thesecurves the residue of anatexis was considered to containclinopyroxene but not plagioclase, in accordance withthe positive Eu anomalies and relatively flat MREE toHREE patterns of the dacitic rocks (Fig. 12). The closecorrespondence between the model partial melting curves
Fig. 18. Al2O3 vs MgO for Washburn rocks. Symbols as in Fig. 9. and the compositions of many of the dacitic rocks in-Filled arrows show effects of mixing between basaltic andesitic magmas dicates that lithologies with compositions similar to Wy-and dacitic magmas on Mount Washburn. Dashed arrow labeled FX oming Province granulite-facies xenoliths are viableshows effects of late-stage fractionation of plagioclase-rich mineral
source rocks. It should be noted, however, that the largeassemblage following mixing for SW Washburn Range dacitic magmas.range in Zr/Th, Sr/Y, and 87Sr/86Sr ratios (Figs 13, 15and 18) and other compositional features of the silicic
magmas were similar in composition to Wyoming Prov- magmas cannot be produced by melting of a single,ince granulite-facies xenoliths. Figure 19 illustrates vari- homogeneous crustal source. Melting of heterogeneousations in Th contents of Washburn dacitic magmas deep crust is required to produce the several distinctrelative to Zr contents and Zr/Th ratios. Also illustrated types of silicic crustal melt present at Washburn volcano,in Fig. 19 are data for Wyoming Province deep-crustal as represented by the diversity of exposed dacitic rocksgranulite-facies xenoliths. The Washburn dacitic rocks (Fig. 11).have low Th contents and a large range in Zr/Th ratios. Alternative sources for the dacitic magmas includePartial melting of Wyoming Province granulite-facies Archean supracrustal amphibolites and granitoids thatlithologies can yield melts with these trace element fea- dominate the exposed basement near Washburn volcano.tures. As an illustration, Fig. 19 shows batch partial The trace element characteristics of the dacitic magmasmelting curves originating from an arbitrarily chosen are not easily reconcilable with partial melting of these
lithologies, however. Most of the supracrustal rocks havecomposition within the fields defined by the xenoliths
694
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Th contents that are too high to serve as appropriate AVP mafic rocks (Fig. 13), is inherited from the mantlesource, or whether it reflects crustal contamination, assources unless some process reduced Th contents in the
partial melts relative to the source lithologies (Fig. 19). the continental crust is also depleted in HFSE relativeto LILE (Weaver & Tarney, 1984; Taylor & McLennan,The most efficient mechanism to lower Th contents is
zircon retention in the source. Retention of zircon (or 1985; Wilson, 1989). This question is particularly relevantin the case of the AVP because the tectonic setting ofany other phase) during melting of the supracrustal
lithologies cannot directly yield magmas similar in com- the field is unclear. Rocks of the AVP were emplacednearly 700 km from the Eocene trench and those whoposition to the Washburn dacitic rocks because the latter
have elevated Zr/Th ratios. Retention of zircon during have advocated a subduction origin for the field havecrustal melting lowers Zr/Th ratios and Th contents of had to invoke complex subduction geometries, includingmelts relative to source rocks (Fig. 19b). In contrast, shallow-angle subduction, imbricate subduction, andWyoming Province granulite-facies xenoliths have low downward buckling of subducted oceanic lithosphereTh contents and an extensive range in Zr/Th ratios that along an axis normal to the trend of the magmatic beltspan those of the Washburn dacitic rocks. (e.g. Coney & Reynolds, 1977; Lipman, 1980; Hum-
In summary, crystal fractionation and crustal con- phreys, 1995).tamination of mafic magmas followed by mixing are To address the problem of whether the elevated LILE/important in controlling compositional trends of the HFSE ratios in Washburn rocks reflect that of the sourceWashburn rocks, but jointly cannot give rise to all of the or crustal contamination, or both, Fig. 20 plots Ba/Nbgeochemical variations observed. Generation of diverse ratios of Washburn rocks vs SiO2 contents (e.g. Nelsonsilicic crustal melts is additionally required to explain the & Davidson, 1993). For the suite as a whole, Ba/Nbtrace element and Sr isotopic characteristics of the dacitic ratios increase as a function of differentiation and com-rocks. The compositional diversity of the dacitic magmas positions of the basaltic andesites project back towardprobably reflects the lithologic heterogeneity and thick- compositions of primitive island arc basalts, althoughness of the Archean Wyoming Province crust and it they greatly exceed those of primitive ocean-island basaltcontrasts with magmatic systems developed in young thin (OIB) and MORB, and OIB-like late Cenozoic basalticcrust, particularly those associated with subduction. In magmas that are common in the nearby Basin and Rangecontinental margin subduction systems such as the south- Province to the south (e.g. Fitton et al., 1988). Moreover,ern Andes, silicic magmas are typically more uniform in bulk Ba and Nb contents for all Washburn rocks plottrace element and isotopic ratios than associated mafic well within the orogenic andesite field of Gill (1981; Fig.magmas (Gerlach et al., 1988; Hickey-Vargas et al., 1989; 20). We interpret these relationships to indicate thatMcMillan et al., 1989; Tormey et al., 1995; Feeley et al., the high Ba/Nb ratios of Washburn mafic rocks are a1998). Although in some cases this results from little characteristic of the mantle source, and that this signaturecrustal involvement following generation of hetero- was amplified in more differentiated rocks by crustalgeneous mantle-derived melts (e.g. Hickey-Vargas et al., contamination. Dudas (1991) and Hooper et al. (1995)1989), to a large extent it is probably a function of the arrived at similar conclusions for Eocene magmatic rockslimited compositional and isotopic contrast between the in the Crazy Mountains and Pacific Northwest, re-parental magmas and the young, subarc crust (Davidson spectively. Subsequent to the study of Hooper et al.et al., 1987; Hildreth & Moorbath, 1988; Feeley, 1993). In (1995), Morris & Hooper (1997) and Morris et al. (2000)contrast, in ancient terrains having experienced complex considered a comparable ‘arc’ signature in the geo-evolutionary histories, including multiple deformation chemistry of rocks from the Colville complex to beevents and magmatic episodes, the crustal column may entirely crustal in origin and therefore unrelated to con-be heterogeneous over relatively small length scales and temporaneous subduction processes. This conclusioncapable of producing at single eruptive centers silicic largely stems from the lack of strong Nb depletions inmelts showing much greater compositional diversity than parental mafic samples present in the complex (e.g. Fig.associated mantle melts. 13c).
The considerations above imply a subduction originor subduction-modified sources for Washburn parental
Mantle source considerations magmas. An interesting feature of AVP volcanic rocksin this regard is that mafic rocks from the shoshoniticDepletions of HFSE and enrichments of LILE relativeSunlight volcano have higher Ba/Nb ratios than Wash-to neighboring elements in diagrams such as Fig. 13 areburn rocks at equivalent bulk compositions (Fig. 20).widely considered diagnostic of magmas derived fromFurther, although Ba contents are significantly elevatedsubduction processes [e.g. Thirlwall et al. (1994) andin Sunlight basalts, Nb contents are comparable withreferences therein]. A critical question is whether thethose in Washburn mafic rocks, implying similar degrees‘arc’ signature present in Washburn trace element pat-
terns, and developed to varying degrees in nearly all of melting in the source. It is unknown if mafic rocks
695
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
in previously metasomatized lithospheric mantle by otherheating or decompression mechanisms proposed forregional magma generation during the mid-Eocene inthe northwestern USA (Dudas, 1991; Morris et al., 2000).
DISCUSSIONSummary: evolution of the Washburnvolcano magmatic systemIntegration of eruptive stratigraphy with petrologic mod-eling has been proven to be a useful technique in un-raveling petrologic processes operating at long-livedcontinental magmatic centers (Dungan et al., 2001). AtWashburn volcano the geochronologic, petrologic, andgeochemical data can be used to construct a model for the
Fig. 20. Ba/Nb vs SiO2 for Washburn rocks. Compositions of arc evolution of the system with the following components:basalts, OIB, and MORB from Pearce (1982). Vectors show expected(1) primitive mantle-derived basaltic magmas frac-trends for mixing with high Ba/Nb crustal melts, fractional crys-
tallization (fc) without alkali feldspar, fractional crystallization with tionated clinopyroxene-rich mineral assemblages and as-alkali feldspar (fc + feldspar), and source modification as a result of similated partial melts of conduit walls at mid- to deep-addition of slab-derived fluids. Shaded field is for Sunlight volcano
crustal levels.basalts (T. C. Feeley, unpublished data, 2001). Inset shows Ba vs Nb(2) Contaminated mafic magmas stalled in the mid-fields for Washburn rocks and Sunlight basalts and illustrates their
affinity to subduction-related andesites. Orogenic andesite and MORB– to deep crust, where they melted rocks with compositionsOIB fields adapted from Gill (1981). similar to Archean granulite-facies rocks and mixed thor-
oughly with the resultant melts. This stage is, of course,analogous to the MASH (mixing, assimilation, storage,from the shoshonitic Independence volcano share the
same characteristics, as Nb data are not available for the homogenization) process of Hildreth & Moorbath (1988)and it produced the dominant volume of andesitic andsuite. The Independence rocks probably have high
Ba/Nb ratios, however, as Ta and Ba contents are basaltic andesitic rocks present on Mt. Washburn andexplains the relative scarcity of either basaltic or daciticdepleted and enriched, respectively, relative to Washburn
basaltic andesites (Fig. 13b). These relationships are rocks.(3) Hybrid magmas ascended to shallow crustalconsistent with variable degrees of source metasomatism
by LILE-rich fluids as suggested by Tatsumi et al. (1986) reservoirs, where they crystallized and in some casesfractionated small amounts of low-pressure mineral as-and summarized by Hawkesworth et al. (1993). As LILE
are particularly mobile in hydrous fluids, especially rel- semblages dominated by plagioclase.(4) Repeated injections of hybrid, intermediate com-ative to HFSE (Tatsumi et al., 1986; Brenan & Watson,
1991), the elevated LILE/HFSE values in mafic rocks position magma into the bases of the shallow chamberscaused minor oscillatory and reverse zoning in any res-from eastern AVP shoshonitic centers are tentatively
consistent with higher degrees of source metasomatism ident phenocrysts.The scenario for rocks in the SW Washburn Rangeby LILE-rich fluids toward the east.
In terms of the tectonic significance of this in- differs from that for rocks on Mt. Washburn in thatthe amount of fractionation of the mafic magmas wasterpretation, an apparent paradox is immediately obvious
if the across-strike K2O enrichment in the AVP in part generally less and the silicic magmas experienced smalldegrees of shallow-level fractionation. Moreover, becausederives from mantle processes involving con-
temporaneous slab dehydration. For example, if the AVP the SW Washburn suite is a bimodal assemblage ofbasaltic andesitic and dacitic rocks it appears that thedoes indeed represent a subduction-related continental
volcanic arc, one might also expect sources further inland extent of homogenization was also less. All of thesefeatures may reflect progressive growth of a crustal mag-to be less affected by metasomatism, as seen in many
modern arcs (e.g. Hickey-Vargas et al., 1989; Davidson matic system as a result of repeated basaltic injections.For example, the composition gap during the early stages& de Silva, 1995). Further, Sr and Nd isotopic com-
positions of all AVP rocks plot well below the field defined of magmatism is interpreted to reflect that mafic magmaswere progressively mixing with silicic crustal melts, butby modern subduction-related basalts as a result of their
low Nd isotopic ratios (Fig. 14). These features indicate that the proportion of silicic melt initially was not greatenough to produce the full spectrum of hybrids. Smallthat although an arc-like component may be present in
AVP magmas, basalt generation may have been triggered proportions of crustally derived melt during the early
696
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
stages of magmatic activity may indicate an initially cold, juxtaposition of compositionally diverse magmas appearsto have occurred in deep chambers repeatedly fluxed bydense crustal column. This process is envisioned as one
in which early batches of mafic magma were emplaced high-temperature, relatively primitive mafic magmas. Insuch environments heating of the crust may be substantial,into the deep crust, but after limited fractionation and
mixing with crustal melts the magmas obtained suffi- resulting in production of partial melts that are effectivelyliquid as a result of nearly complete resorption of anyciently low densities to promote further ascent. The
process of basalt emplacement and crystallization oc- entrained pre-existing crystals (Watson, 1982; Tsuchi-yama, 1985; Huppert & Sparks, 1988; Koyaguchi &curred repeatedly, leading to large-scale heating of the
crust. Higher temperatures in the crust rapidly increased Kaneko, 1999). Furthermore, as proposed by Hildreth& Moorbath (1988) and suggested by the stratigraphicthe proportion of silicic melt (Huppert & Sparks, 1988),
decreased the density of the crustal column, and increased succession documented in this paper, mixing in deep-crustal magma chambers may be incremental and involvethe residence time for basaltic magmas. Collectively,
these processes induced greater degrees of high-pressure repeated blending of fractionated mantle-derived maficmagmas, silicic crustal melts, and evolved magmas whichfractionation, contamination, and homogenization with
silicic crustal melts. At all temporal stages in the evolution themselves are hybrids formed during earlier injection–mixing episodes. In this manner, endmember magmasof Washburn volcano the hybrid magmas stalled in
shallow magma chambers before eruption. This episode progressively converge toward intermediate compositionsthat can readily homogenize. Subsequent ascent andis documented by growth of new, plagioclase-rich mineral
assemblages in equilibrium with the hybrid liquids. In crystallization of hybrid magmas in shallow chamberswill overprint these earlier mixing episodes.the case of the SW Washburn system where early batches
of silicic magma were probably small and the crust In contrast to the Washburn system, many calc-alkalinesuites contain dramatic mineralogic and petrographicrelatively cold, convective velocities within the shallow
chambers may have been sufficiently low to permit frac- evidence for mixing, particularly where homogenizationis incomplete and compositionally banded tephra ortionation of resident phenocrysts.commingled mafic inclusions occur within more silicichosts (Wilcox, 1999). In these systems magma interaction
Petrologic significance appears to have occurred rapidly during injection ofIn light of the foregoing discussion one of the more basaltic magma into the bases of shallow silicic chambersintriguing aspects of the Washburn suite is that although (Blake et al., 1965; Eichelberger, 1975; Anderson, 1976;bulk compositions strongly suggest that the intermediate Bacon, 1986; Coombs et al., 2000). There is also abundantcomposition rocks are hybrids, compositional and modal evidence in these systems that basaltic injection is re-data for phenocrysts show only limited evidence for sponsible for triggering eruption, thereby arresting re-mixing in the form of mineral–melt disequilibria or equilibration of high- and low-temperature phenocrystcoexisting high- and low-temperature phenocryst as- assemblages within the new hybrid magma (Pallister etsemblages (e.g. Figs 5–8). In this regard the rocks resemble al., 1992; Feeley & Sharp, 1996). At Washburn volcanothe ‘cryptic hybrids’ of Dungan (1987) and Kerr et al. injections into shallow chambers may also have triggered(1999). Dungan (1987) suggested that the most likely eruptions. These injections, however, were probably hy-condition for generating cryptic hybrids involves blending brid intermediate magmas compositionally similar toof mineralogically similar, low-viscosity tholeiitic mafic resident magmas, thereby producing little discerniblemagmas, although he also proposed a more general textural or mineralogical disequilibrium in the eruptedmodel applicable to calc-alkaline and bimodal systems lavas.where endmember magmas differ markedly in solidustemperatures, densities, viscosities, and mineral as-
Implications for magmagenesis in the AVPsemblages. Important elements of the model includeincremental mixing of superheated, crystal-poor magmas Recognition of the Washburn magmas as cryptic hybrids
provides new insight into magmatic processes operatinglong before eruption. Although Dungan (1987) did notfurther identify specific magmatic environments where in the AVP. An outstanding question concerning the
origin of magmatic rocks in the AVP is the extent tothese conditions are optimized, he noted that the mainrequirement is that sufficient time must exist following which western belt calc-alkaline magmas interacted with
the continental crust. At other magmatic fields in themixing for the hybrid magmas to achieve textural andmineralogical equilibrium before eruption. ‘Challis arc’ (Fig. 1), a magma mixing origin has been
demonstrated for calc-alkaline rocks (e.g. Morris &The petrogenetic model developed in this paper impliesthat generation of cryptic hybrids in compositionally Hooper, 1997; Morris et al., 2000). However, before this
study, a simple closed-system history was assumed for thediverse calc-alkaline systems may be facilitated in deep-crustal zones of differentiation. At Washburn volcano petrogenesis of the AVP calc-alkaline magmas because
697
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
no convincing evidence had been presented to indicate shoshonitic differentiation trends in the AVP. The com-positional relationships among rocks at Washburn vol-crustal involvement during differentiation. Indeed, the
simple phenocryst assemblages of the rocks were cited cano indicate, however, that AVP calc-alkaline andesiticmagmas evolved mainly by magma mixing and thatas strong evidence for differentiation by closed-system
fractional crystallization. This facet was established over differentiation of parental magmas involved fractionalcrystallization plus one or more open-system processesa century ago in the classic study of Iddings (1891) on
AVP calc-alkaline rocks exposed at the Electric Peak– such as mixing and assimilation of crust. All Washburnrocks thus appear to contain a crustal component andSepulcher Mountain eruptive center (Fig. 1), whichnone can reasonably be inferred to be direct differentiatesproduced rocks temporally, compositionally, and petro-of primary mantle-derived magmas, as has been suggestedgraphically similar to those at Washburn volcano (Lindsayfor magmas parental to shoshonites at Independence& Feeley, 1999). At this center Iddings (1891) utilizedvolcano (e.g. Fig. 10; Meen, 1987b; Meen & Eggler,optical and chemical techniques available at the time to1987). Differences in the location of crustal partial meltingargue for textural and compositional equilibrium betweenor sources during silicic melt generation also contributedphenocrysts and whole rocks, and correlated changes into the development of compositional diversity at Wash-phenocryst modes with differentiation nearly identical toburn volcano. Given these results, magma mixing, partialthose documented here. In his view these features sug-melting of the crust, and contamination of parentalgested that: ‘the chemical differences of igneous rocksmagmas must therefore be considered as important inare the result of a chemical differentiation of a generaleither generating or influencing differences between calc-magma’ (Iddings, 1891). Subsequent petrologic studiesalkaline and shoshonitic suites in the AVP. In this regardbased on limited chemical datasets maintained this opin-we do not dismiss crystal fractionation as trivial, but theion and further argued that differentiation of AVP calc-variation from rocks that follow calc-alkaline trends inalkaline magmas involved simple fractional crystallizationthe west to those that follow shoshonitic trends in theof plagioclase-rich assemblages (Schultz, 1962; Petermaneast cannot simply reflect higher pressures of fractionationet al., 1970; Love et al., 1976; LaPointe, 1977).to the east in Moho-level magma chambers in the absenceThe significance of the open-system petrogenetic modelof crustal interaction. Although the origin of the across-advocated here lies not only in the new insight it providesstrike K2O trend is still unclear, further progress intointo magma generation processes in the AVP, but perhapsbetter understanding the roles of open- vs closed-systemmore importantly in the constraints it places on mech-processes in the AVP is anticipated with additional geo-anisms responsible for producing calc-alkaline suites inchemical studies of individual centers.the western belt and shoshonitic suites in the eastern belt.
Meen (1987b) used experimental data for shoshoniticrocks at the eastern belt Independence volcano to arguethat parental magmas for both suites may be produced
CONCLUSIONSfrom mantle-derived subalkaline basaltic magmas withoutinteraction with crustal rocks. In this scenario, shoshonitic (1) Magmatic rocks at Washburn volcano range in age
from >55 to 52 Ma and form a medium- to high-K,differentiation trends result from high-pressure (10 kbar)fractional crystallization of subalkaline basaltic magmas, calc-alkaline suite ranging in composition from basaltic
andesitic through dacitic.whereas calc-alkaline magma series are generated fromshallow pressure, anhydrous fractional crystallization in- (2) Mineral textural and composition features along
with whole-rock compositional data can be interpretedvolving significant olivine in the crystallizing assemblage.The effect of the latter process is to rapidly enrich SiO2 as reflecting near-equilibrium crystallization of observed
phenocryst phases.contents and deplete MgO contents relative to K2O andproduce relatively linear, shallow arrays comparable with (3) Stratigraphic relationships, along with petrogenetic
modeling of trace element and Sr and Nd isotopic data,those in Figs 9 and 10b for the Washburn suite. Lendinguncertainty to this interpretation, however, is the re- demonstrate that the Washburn rocks were produced by
progressive mixing of variably fractionated and con-cognition that the calc-alkaline trend may also be pro-duced by crustal contamination of anhydrous mafic taminated mantle-derived melts and heterogeneous silicic
crustal melts.magmas evolving along high-pressure liquid lines of des-cent [see fig. 11 of Meen (1987b)]. (4) Nd and Sr isotopic compositions along with trace
element data indicate that silicic melts in the WashburnAs outlined above, before this study no convincingpetrologic evidence existed to indicate that crustal in- system are derived from deep-crustal rocks broadly similar
in composition to granulite-facies xenoliths in the Wy-volvement was important in the generation of magmasat calc-alkaline centers in the AVP, allowing for a simple oming Province. Data for basaltic andesitic rocks indicate
that the ‘subduction’ signature is of mantle origin, al-‘increasing depth to magma chamber relationship’ toexplain the eastward transition from calc-alkaline to though it may not be related to active subduction.
698
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
Brophy, J. G. (1991). Composition gaps, critical crystallinity, and(5) Compositional and petrographic diversity in thefractional crystallization in orogenic (calc-alkaline) magmatic systems.Washburn suite were produced by homogenization ofContributions to Mineralogy and Petrology 109, 173–182.mafic and silicic crustal melts in deep-crustal magma
Brown, T. M. (1982). Geology, paleontology, and correlation of Eocenechambers followed by ascent and growth of new plagio- volcaniclastic rocks, southeast Absaroka Range, Hot Springs County,clase-rich mineral assemblages in equilibrium with the Wyoming. US Geological Survey Professional Paper 1201-A, 1–75.hybrid liquids. Many calc-alkaline magmas may therefore Burchfield, B. C., Cowan, D. S. & Davis, G. A. (1992). Tectonic
overview of the Cordilleran orogen in the western United States.be cryptic hybrids recognizable only on chemical grounds,In: Burchfield, B. C., Lipman, P. W. & Zoback, M. L. (eds) Theparticularly in systems where mixing precedes and isCordilleran Orogen: Conterminous U.S. Geological Society of America, Geologywidely separated from crystallization in space and time.of North America G-3, 407–479.(6) Given the role ascribed to crustal processes at Chadwick, R. A. (1970). Belts of eruptive centers in the Absaroka–
Washburn volcano, partial melting of the crust, crustal Gallatin Volcanic Province, Wyoming–Montana. Geological Society ofcontamination of parental magmas, and magma mixing America Bulletin 81, 267–274.
Chadwick, R. A. (1985). Overview of Cenozoic volcanism in themust play some role in either generating or influencingwest–central United States. In: Kaplan, S. S. & Flores, R. M. (eds)across-strike K2O trends in the AVP. The variationRocky Mountain SEPM Cenozoic Paleography of the West–Central Unitedbetween rocks that follow calc-alkaline trends in the westStates Symposium 3, 359–381.to those that follow shoshonitic trends in the east cannot
Clemens, J. D. (1989). The importance of residual source materialsimply reflect higher pressures of fractionation to the east (restite) in granite petrogenesis: a comment. Journal of Petrology 30,in Moho-level magma chambers in the absence of crustal 1313–1316.interaction. Collerson, K. D., Hearn, B. C., MacDonald, R. A., Upton, B. G. J.
& Harmon, R. S. (1989). Composition and evolution of lowercontinental crust: evidence from xenoliths in Eocene lavas from theBearpaw Mountains, Montana. New Mexico Bureau of Mines and Mineral
Resources Bulletin 131, 57.ACKNOWLEDGEMENTSConey, P. J. & Reynolds, S. J. (1977). Cordilleran Benioff zones. NatureThis work was supported by NSF grant EAR-9725287
270, 403–406.to T.C.F. and GSA Penrose and Sigma Xi grants to Coombs, M. L., Eichelberger, J. C. & Rutherford, M. J. (2000). MagmaC.R.L. The Radiation Center at Oregon State University storage and mixing conditions for the 1953–1974 eruptions ofprovided the instrumental neutron activation analyses. Southwest Trident volcano, Katmai National Park, Alaska. Con-
tributions to Mineralogy and Petrology 140, 99–118.The authors thank Francois Bussy for assistance with theDalrymple, G. B., Alexander, E. C., Lanphere, M. A. & Kraker, G.microprobe, Megan O’Connor for able and cheerful field
P. (1981). Irradiation of samples for 40Ar–39Ar dating using theassistance, and Yellowstone National Park personnel forGeological Survey TRIGA reactor. US Geological Survey, Professionalpermission to obtain samples from the field area andPaper, 1176, 1–55.
hospitality. Discussions with Margaret Hiza helped to Davidson, J. P. & de Silva, S. L. (1995). Late Cenozoic magmatism ofstimulate many of the ideas presented in this paper. the Bolivian Altiplano. Contributions to Mineralogy and Petrology 119,
387–408.Helpful reviews by Peter Hooper, Andrew Kerr, GeorgeDavidson, J. P., Dungan, M. A., Ferguson, K. M. & Colucci, M. T.Morris, Marjorie Wilson, and Jim Meen on earlier ver-
(1987). Crust–magma interactions and the evolution of arc magmas:sions of this paper are appreciated, as is the editorialthe San Pedro–Pellado volcanic complex, southern Chilean Andes.work of Marjorie Wilson and Alastair Lumsden.Geology 15, 443–446.
DeBari, S. M. & Sleep, N. H. (1991). High-Mg, low Al-bulk compositionof the Talkeetna island arc, Alaska: implications for primary magmasand the nature of the arc crust. Geological Society of America BulletinREFERENCES 103, 37–47.
Anders, E. & Ebihara, M. (1982). Solar-system abundances of the Decker, P. L. (1990). Style and mechanics of liquefaction-relatedelements. Geochimica et Cosmochimica Acta 46, 2363–2380. deformation lower Absaroka Volcanic Supergroup (Eocene), Wy-
Anderson, A. T. (1976). Magma mixing: petrological process and oming. Geological Society of America, Special Paper 240, 1–71.volcanological tool. Journal of Volcanology and Geothermal Research 1, DePaolo, D. J. (1981). Trace element and isotopic effects of combined3–33. wallrock assimilation and fractional crystallization. Earth and Planetary
Armstrong, R. L. (1978). Cenozoic igneous history of the U.S. Cordillera Science Letters 53, 189–202.from lat 42° to 49°N. Geological Society of America Memoir 42, 265–282. Dickinson, W. R. & Hatherton, T. (1967). Andesitic volcanism and
Bacon, C. R. (1986). Magmatic inclusions in silicic and intermediate seismicity around the Pacific. Science 157, 801–803.volcanic rocks. Journal of Geophysical Research 91, 6091–6112. Dorf, E. (1964). The petrified forests of Yellowstone National Park.
Blake, D. H., Elwell, R. W. D., Gibson, I. L., Skelhorn, R. R. & Scientific American 210, 106–112.Walker, G. P. L. (1965). Some relationships resulting from the Dostal, J., Robichaud, D. A., Church, B. N. & Reynolds, P. H. (1998).intimate association of acid and basic magmas. Journal of the Geological Eocene Challis–Kamloops volcanism in central British Columbia:Society, London 121, 31–49. an example from the Buck Creek Basin. Canadian Journal of Earth
Brenan, J. M. & Watson, E. B. (1991). Partitioning of trace elements Sciences 35, 951–963.between olivine and aqueous fluids at high P–T conditions: im- Dudas, F. O. (1991). Geochemistry of igneous rocks from the Crazyplications for the effect of fluid composition on trace-element trans- Mountains, Montana, and tectonic models for the Montana alkalic
province. Journal of Geophysical Research 96, 13261–13277.port. Earth and Planetary Science Letters 107, 672–688.
699
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
Dudas, F. O., Carlson, R. W. & Eggler, D. H. (1987). Regional Middle Gerlach, D. C., Frey, F. A., Moreno-Roa, H. & Lopez-Escobar, L.(1988). Recent volcanics from the Puyehue–Cordon Caulle region,Proterozoic enrichment of the subcontinental mantle source ofsouthern Andes, Chile (40°5′S). Journal of Petrology 29, 333–382.igneous rocks from central Montana. Geology 15, 22–25.
Gill, J. (1981). Orogenic Andesites and Plate Tectonics. Berlin: Springer, 390Dungan, M. A. (1987). Open system magmatic evolution of the Taospp.Plateau volcanic field, northern New Mexico: II. The genesis of
Grove, T. L. & Baker, M. B. (1984). Phase equilibrium controls oncryptic hybrids. Journal of Petrology 28, 955–977.the tholeiitic versus calc-alkaline differentiation trends. Journal ofDungan, M. A., Wulff, A. & Thompson, R. (2001). Eruptive stratigraphyGeophysical Research 89, 3253–3274.of the Tatara–San Pedro Complex, 36°S, Southern Volcanic Zone,
Grove, T. L. & Donnelly-Nolan, J. M. (1986). The evolution of silicicChilean Andes: reconstruction method and implications for magmalavas at Medicine Lake Volcano, California: implications for theevolution at long-lived arc volcanic centers. Journal of Petrology 42,origin of compositional gaps in calc-alkaline series lavas. Contributions555–626.to Mineralogy and Petrology 92, 281–302.Dutch, S. I. & Nielsen, P. A. (1990). The Archean Wyoming Province
Harlan, S. S., Snee, L. W. & Geissman, J. W. (1996). 40Ar/39Arand its relations with adjacent Proterozoic provinces. In: Lewry, J.geochronology and paleomagnetism of Independence volcano, Ab-F. & Stauffer, M. R. (eds) The Early Proterozoic Trans-Hudson Orogen ofsaroka Volcanic Supergroup, Beartooth Mountains, Montana. Ca-North America. Geological Society of Canada Special Paper 37, 287–300.nadian Journal of Earth Sciences 33, 1648–1654.Eaton, J. G. (1982). Paleontology and correlation of Eocene volcanic
Hauge, T. A. (1985). Gravity-spreading origin of the Heart Mountainrocks in the Carter Mountain area, Park County, southeasternallochthon, northwestern Wyoming. Geological Society of America BulletinAbsaroka Range, Wyoming. University of Wyoming Contributions to96, 1440–1456.Geology 21, 1019–1022.
Hauri, E. H., Wagner, T. P. & Grove, T. L. (1994). Experimental andEggler, D. H. (1972). Amphibole stability in H2O-undersaturated calc-natural partitioning of Th, U, Pb, and other trace elements betweenalkaline melts. Earth and Planetary Science Letters 15, 28–34.garnet, clinopyroxene, and basaltic melts. Chemical Geology 117, 149–Eichelberger, J. C. (1975). Origin of andesite and dacite; evidence of166.magma mixing at Glass Mountain in California and other Circum-
Hawkesworth, C. J., Gallagher, K., Hergt, J. M. & McDermott, F.Pacific volcanoes. Geological Society of America Bulletin 86, 1381–1391.(1993). Mantle and slab contributions in arc magmas. Annual ReviewFeeley, T. C. (1993). Crustal modification during subduction zoneof Earth and Planetary Sciences 21, 175–204.magmatism: evidence from the southern Salar de Uyuni region
Hawthorne, F. C. (1981). Crystal chemistry of amphiboles. In: Veblin,(20°–22°S), Central Andes. Geology 21, 1019–1022.
D. R. (ed.) Amphiboles and Other Hydrous Pyriboles—Mineralogy. Min-Feeley, T. C. & Davidson, J. P. (1993). Petrology of calc-alkaline lavas
eralogical Society of America, Reviews in Mineralogy 9A, 1–102.at volcan Ollague and the origin of compositional diversity at central
Heltz, R. T. (1973). Phase relations of basalts in their melting rangeAndean stratovolcanoes. Journal of Petrology 35, 1295–1340.
at PH2O= 5 kb as a function of oxygen fugacities. Journal of PetrologyFeeley, T. C. & Dungan, M. A. (1996). Compositional and dynamic
14, 443–481.controls on mafic–silicic magma interactions at continental arc Hickey-Vargas, R., Moreno, H., Lopez-Escobar, L. & Frey, F. A.volcanoes: evidence from Cordon El Guadal, Tatara–San Pedro (1989). Geochemical variations in Andean basaltic and silicic lavasComplex, Chile. Journal of Petrology 37, 1547–1577. from the Villarrica–Lanin volcanic chain (39·5°S): an evaluation of
Feeley, T. C. & Sharp, Z. D. (1996). Chemical and hydrogen isotope source heterogeneity, fractional crystallization, and crustal as-evidence for in situ dehydrogenation of biotite in silicic magma similation. Contributions to Mineralogy and Petrology 103, 261–386.chambers. Geology 24, 1021–1024. Hildreth, W. & Moorbath, S. (1988). Crustal contributions to arc
Feeley, T. C., Dungan, M. A. & Frey, F. A. (1998). Geochemical magmatism in the Andes of central Chile. Contributions to Mineralogyconstraints on the origin of mafic and silicic magmas at Cordon El and Petrology 98, 455–499.Guadal, Tatara–San Pedro Complex, central Chile. Contributions to Hiza, M. M. (1999). The geochemistry and geochronology of the EoceneMineralogy and Petrology 131, 393–411. Absaroka volcanic province, northern Wyoming and southwest
Feeley, T. C., Lindsay, C. R. & Cosca, M. A. (1999). Timing and Montana, USA. Ph.D. thesis, Oregon State University, Corvallis.compositional evolution of magmatism at the Sunlight volcano, Holder, R. W. & Holder, G. A. M. (1988). The Colville Batholith:Absaroka volcanic province, Wyoming. Geological Society of America, Tertiary plutonism in northeast Washington associated with grabenAbstracts with Programs 31, A-165. and core-complex (gneiss dome) formation. Geological Society of America
Fitton, J. G., James, D., Kempton, P. D., Ormerod, D. S. & Leeman, Bulletin 100, 1971–1980.W. P. (1988). The role of lithospheric mantle in the generation of Hooper, P. R., Bailey, D. G. & McCarley Holder, G. A. (1995). Tertiarylate Cenozoic basic magmas in the western United States. Journal of calc-alkaline magmatism associated with lithospheric extension inPetrology, Special Lithosphere Issue, 331–349. the Pacific Northwest. Journal of Geophysical Research 100, 10303–
Foden, J. D. (1983). The petrology of the calkalkaline lavas of Rindjani 10319.volcano, East Sunda arc: a model for arc petrogenesis. Journal of Huebner, J. S. & Sato, M. (1970). The oxygen fugacity relationshipsPetrology 24, 98–130. of manganese oxide and nickel oxide buffers. American Mineralogist
Foden, J. D. & Green, D. H. (1992). Possible role of amphibole in 55, 934–952.the origin of andesite: some experimental and natural evidence. Humphreys, E. D. (1995). Post-Laramide removal of the Farallon slab,Contributions to Mineralogy and Petrology 109, 479–493. western United States. Geology 23, 987–990.
Fraser, K. J., Hawkesworth, C. J., Erlank, A. J., Mitchell, R. H. & Huppert, H. E. & Sparks, R. S. J. (1988). The generation of graniticScott-Smith, B. H. (1985). Sr, Nd, and Pb isotope and minor element magma by intrusion of basalt into continental crust. Journal of Petrologygeochemistry of lamproites and kimberlites. Earth and Planetary Science 29, 599–624.Letters 76, 57–70. Iddings, J. P. (1891). The eruptive rocks of Electric Peak and Sepulchre
Frey, F. A., Green, D. H. & Roy, S. D. (1978). Integrated models of Mountain. US Geological Survey 12th Annual Report 577–664.basalt petrogenesis: a study of quartz tholeiites to olivine melilitites Irving, A. J. & Frey, F. A. (1978). Trace element abundances in
megacrysts and their host basalts: constraints on partition coefficientsfrom south eastern Australia. Journal of Petrology 19, 463–513.
700
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
and megacryst genesis. Geochimica et Cosmochimica Acta 48, 1201–1221. Wyoming and Montana. Geological Society of America Bulletin 87,1455–1462.Janecke, S. U. & Snee, L. W. (1993). Timing and episodicity of middle
Eocene volcanism and onset of conglomerate deposition, Idaho. Luedke, R. G. (1994). Map showing the distribution, composition, andage of early and middle Cenozoic volcanic centers in Idaho, Mon-Journal of Geology 101, 603–621.
Johnson, D. M., Hooper, P. R. & Conrey, R. M. (1999). XRF analysis tana, west–central South Dakota, and Wyoming. US Geological Survey
Map I-2291-C.of rocks and minerals for major and trace elements on a single lowdilution Li-tetraborate fused bead. Advances in XRF Analysis 41, MacDonald, R., Upton, B. G. J., Collerson, K. D., Hearn, B. C., Jr
& James, D. (1992). Potassic mafic lavas of the Bearpaw Mountains,843–867.Johnson, K. T. M. (1998). Experimental determination of partition Montana: mineralogy, chemistry, and origin. Journal of Petrology 33,
305–346.coefficients for rare earth and high-field-strength elements betweenclinopyroxene, garnet, and basaltic melt at high pressures. Con- McMillan, N. F. & Dungan, M. A. (1984). Magma mixing in the Taos
Plateau magmatic system. In: Dungan, M. A., Grove, T. L. &tributions to Mineralogy and Petrology 133, 60–68.Joswiak, D. (1992). Composition and evolution of the lower crust, Hildreth, W. (eds) Proceedings of the ISEM Field Conference on Open
Magmatic Systems. Dallas, TX: Institute for the Study of Earth andcentral Montana: evidence from granulite xenoliths. M.S. thesis,University of Washington, Seattle. Man, pp. 110–112.
McMillan, N. F. & Dungan, M. A. (1986). Magma mixing as aKennedy, A. K., Grove, T. L. & Johnson, R. W. (1990). Experimentaland major element constraints on the evolution of lavas from Lihir petrogenetic process in the development of the Taos Plateau volcanic
field, New Mexico. Journal of Geophysical Research 91, 6029–6045.Island, Papua New Guinea. Contributions to Mineralogy and Petrology
104, 722–734. McMillan, N. J., Harmon, R. S., Moorbath, S., Lopez-Escobar, L. &Strong, D. F. (1989). Crustal sources involved in continental arcKerr, A. C., Kent, R. W., Thomson, B. A., Seedhouse, J. K. &
Donaldson, C. H. (1999). Geochemical evolution of the Tertiary magmatism: a case study of volcan Mocho–Choshuenco, southernChile. Geology 17, 1152–1156.Mull volcano, western Scotland. Journal of Petrology 40, 873–908.
Kilinc, A., Carmichael, I. S. E., River, M. L. & Sack, R. O. (1983). Meen, J. K. (1985). The origin and evolution of a continental volcano—Independence, Montana. Ph.D. thesis, Pennsylvania State Uni-The ferric–ferrous ratio of natural silicate liquids equilibrated in air.
Contributions to Mineralogy and Petrology 83, 136–140. versity, University Park.Meen, J. K. (1987a). Sr, Nd, Pb, isotopic compositions of ArcheanKoyaguchi, T. & Kaneko, K. (1999). A two-stage thermal evolution
model of magmas in continental crust. Journal of Petrology 40, 241–254. basement rocks, Boulder River region, Beartooth Mountains, Mon-tana. Isochron West 50, 13–24.Langmuir, C. H., Vocke, R. D., Jr, Hanson, G. H. & Hart, S. R.
(1978). A general mixing equation with applications to Icelandic Meen, J. K. (1987b). Formation of shoshonites from calcalkaline basaltmagmas: geochemical and experimental constraints from the typebasalts. Earth and Planetary Science Letters 37, 380–392.
LaPointe, D. D. (1977). Geology of a portion of the Eocene Sylvan locality. Contributions to Mineralogy and Petrology 104, 309–331.Meen, J. K. & Eggler, D. H. (1987). Petrology and geochemistry ofPass volcanic center, Absaroka Range, Wyoming. M.S. thesis, Uni-
versity of Montana, Missoula, 60 pp. the Cretaceous Independence volcanic suite, Absaroka Mountains:clues to the composition of the Archean sub-Montanan mantle.Laul, J. C. (1979). Neutron activation analysis of geologic materials.
Atomic Energy Reviews 17, 603–695. Geological Society of America Bulletin 98, 238–247.Meen, J. K. & Eggler, D. H. (1989). Chemical and isotopic compositionsLeake, B. E. (1978). Nomenclature of amphiboles. American Mineralogist
63, 1023–1053. of Absaroka granitoids, southwestern Montana; evidence for deep-seated Archean amphibolite basement in the Beartooth region.Leeman, W. P., Menzies, M. A., Mattey, D. J. & Embree, G. F. (1985).
Sr, Nd, and Pb isotopic composition of deep crustal xenoliths from Contributions to Mineralogy and Petrology 102, 462–477.Morris, G. A. & Hooper, P. R. (1997). Petrogenesis of the Colvillethe Snake River Plain: evidence for Archean basement. Earth and
Planetary Science Letters 75, 354–368. igneous complex, northeast Washington: implications for Eocenetectonics in the northern U.S. Cordillera. Geology 25, 831–834.Le Maitre, R. W. (1989). A Classification of Igneous Rocks and a Glossary of
Terms. Recommendations of the International Union of Geological Sciences Morris, G. A., Larson, P. B. & Hooper, P. R. (2000). ‘Subductionstyle’ magmatism in a non-subduction setting: the Colville IgneousSubcommission on the Systematics of Igneous Rocks. Oxford: Blackwell.
Lindsay, C. R. & Feeley, T. C. (1999). Petrology and geochemistry of complex, NE Washington State, USA. Journal of Petrology 41, 43–67.Mueller, P. A., Wooden, J. L., Odom, A. L. & Bowes, D. R. (1982).the Electric Peak–Sepulcher Mountain eruptive center: implications
for the petrogenesis of calc-alkaline magmas in the Absaroka volcanic Geochemistry of the Archean rocks of the Quad Creek and Hellroar-ing Plateau areas of the eastern Beartooth Mountains. Montana Bureauprovince. Geological Society of America, Abstracts with Programs 31, A-353.
Lipman, P. W. (1980). Cenozoic volcanism in the western United of Mines and Geology Special Publication 84, 69–82.Mueller, P. A., Wooden, J. L., Schulz, K. & Bowes, D. R. (1983).States: implications for continental tectonics. In: Burchfield, J. C.,
Oliver, J. E. & Silver, L. T. (eds) Studies in Geophysics, Continental Incompatible-element-rich andesitic amphibolites from the Archeanof Montana and Wyoming: evidence for mantle metasomatism.Tectonics. Washington, DC: National Academy of Sciences, pp. 161–
174. Geology 11, 203–206.Mueller, P. A., Wooden, J. L., Henry, D. J. & Bowes, D. R. (1985).Lipman, P. W., Prostka, H. J. & Christiansen, R. L. (1972). Cenozoic
volcanism and plate-tectonic evolution of the western United States. Archean crustal evolution of the eastern Beartooth Mountains,Montana and Wyoming. Montana Bureau of Mines and Geology SpecialPhilosophical Transactions of the Royal Society of London, Series A 271,
217–248. Publication 92, 9–20.Nelson, S. T. & Davidson, J. P. (1993). Interactions between mantle-Love, J. D., Tschudy, R. H. & Obradovich, J. D. (1975). Dating a
Laramide orogeny, northwestern Wyoming. US Geological Survey derived magmas and mafic crust, Henry Mountains, Utah. Journal
of Geophysical Research 98, 1837–1852.Professional Paper 975, 1–51.Love, L. L., Kudo, A. M. & Love, D. W. (1976). Dacites of Bunsen Norman, M. C. & Mertzman, S. A. (1991). Petrogenesis of Challis
volcanics from central and southwestern Idaho: trace element andPeak, the Birch Hills, and the Washakie Needles, northwesternWyoming, and their relationship to the Absaroka volcanic field, Pb isotopic evidence. Journal of Geophysical Research 96, 13279–13293.
701
JOURNAL OF PETROLOGY VOLUME 43 NUMBER 4 APRIL 2002
O’Brien, H. E., Irving, A. J., McCallum, I. S. & Thirlwall, M. F. evidence from mineral chemistry. Journal of Volcanology and Geothermal
Research 74, 275–295.(1995). Strontium, neodymium, and lead isotopic evidence for theSundell, K. A. & Eaton, J. G. (1982). Stratigraphic relations within theinteraction of post-subduction asthenospheric potassic mafic magmas
southeastern Absaroka volcanic sequence, northwestern Wyoming.of the Highwood Mountain, Montana, USA, with ancient WyomingWyoming Geological Association Annual Field Conference Guidebook 34, 65–71.craton lithospheric mantle. Geochimica et Cosmochimica Acta 59, 4539–
Tatsumi, Y., Hamilton, D. L. & Nesbitt, R. W. (1986). Chemical4556.characteristics of fluid phase released from a subducted lithospherePallister, J. S., Hoblitt, R. P. & Reyes, A. G. (1992). A basalt triggerand the origin of arc magmas: evidence from high-pressure ex-for the 1991 eruptions of Pinatubo volcano? Nature 356, 426–428.periments and natural rocks. Journal of Volcanology and GeothermalPeacock, M. A. (1931). Classification of igneous rocks. Journal of GeologyResearch 29, 293–309.39, 54–67.
Taylor, H. P., Jr (1980). The effects of assimilation of country rocksPearce, J. A. (1982). Trace element characteristics of lavas fromby magmas on 18O/16O and 87Sr/86Sr systematics in igneous rocks.destructive plate boundaries. In: Thorpe, R. S. (ed.) Andesites. NewEarth and Planetary Science Letters 47, 243–254.York: John Wiley, pp. 525–548.
Taylor, S. R. & McLennan, S. M. (1985). The Continental Crust: itsPeccerillo, A. & Taylor, S. R. (1976). Geochemistry of Eocene calc-Composition and Evolution. Oxford: Blackwell Scientific, 312 pp.alkaline volcanic rocks in the Kastamonu area, northern Turkey.
Thirlwall, M. F., Smith, T. E., Graham, A. M., Theodorou, N.,Contributions to Mineralogy and Petrology 58, 63–81.Hollings, P., Davidson, J. P. & Arculus, R. J. (1994). High fieldPeterman, Z. E., Doe, B. R. & Prostka, H. J. (1970). Lead andstrength element anomalies in arc lavas: source or process? Journalstrontium isotopes in rocks of the Absaroka volcanic field, Wyoming.of Petrology 35, 819–838.Contributions to Mineralogy and Petrology 27, 121–130.
Tormey, D. R., Frey, F. A. & Lopez-Escobar, L. (1995). GeochemistryProdehl, C. & Lipman, P. W. (1989). Crustal structure of the Rockyof the active Azufre–Planchon–Peteroa volcanic complex, ChileMountain region. Geological Society of America Memoir 172, 249–284.(35°15′S): evidence for multiple sources and processes in a Cor-Prostka, H. J., Blank, H. R., Jr, Christiansen, R. L. & Ruppel, E. T.dilleran arc magmatic system. Journal of Petrology 36, 265–298.(1975). Geologic map of the Tower Junction Quadrangle, Yel-
Tsuchiyama, A. (1985). Dissolution kinetics of plagioclase in the meltlowstone National Park, Wyoming and Montana. US Geological Surveyof the system diopside–albite–anorthite, and origin of dusty plagio-Map GQ-1247.clase in andesites. Contributions to Mineralogy and Petrology 89, 1–16.Reiners, P. W., Nelson, B. K. & Ghiorso, M. S. (1995). Assimilation
Wagner, T. P., Donnelly-Nolan, J. M. & Grove, T. L. (1995). Evidenceof felsic crust by basaltic magmas: thermal limits and extents of crustalof hydrous differentiation and crystal accumulation in the low-MgO,contamination of mantle-derived magmas. Geology 23, 563–566.high Al2O3 Lake Basalt from Medicine Lake volcano, California.Roeder, R. L. & Emslie, R. F. (1970). Olivine–liquid equilibrium.Contributions to Mineralogy and Petrology 121, 210–216.
Contributions to Mineralogy and Petrology 29, 275–289.Watson, E. B. (1982). Basalt contamination by continental crust: someRollinson, H. (1993). Using Geochemical Data: Evaluation, Presentation,
experiments and models. Contributions to Mineralogy and Petrology 80,Interpretation. Harlow: Longman, 352 pp.
73–87.Ruppel, E. T. (1972). Geology of pre-Tertiary rocks in the northernWeaver, B. L. & Tarney, J. (1984). Empirical approach to estimating
part of Yellowstone National Park, Wyoming. US Geological Surveythe composition of the continental crust. Nature 310, 375–377.
Professional Paper 729-A, 1–66.Wells, P. A. (1977). Pyroxene thermometry in simple and complex
Samson, S. D. & Alexander, E. C. (1987). Calibration of the in-systems. Contributions to Mineralogy and Petrology 62, 129–139.
terlaboratory 40Ar/39Ar dating standard, MMhb-1. Chemical Geology Wheeler, J. O., Brookfield, A. J., Gabrielse, H., Monger, J. W., Tipper,66, 27–34. H. W. & Woodsworth, G. J. (1991). Terrane map of the Canadian
Sato, H. (1977). Nickel content in basaltic magmas: identification ofCordillera. Geological Survey of Canada, Map 1713A, scale
primary magmas and a measure of the degree of olivine fractionation. 1:2 000 000.Lithos 10, 113–120. White, A J. R. & Chappell, B. W. (1977). Ultrametamorphism and
Scambos, T. A. (1991). Isotopic and trace-element characteristics of granitoid genesis. Tectonophysics 43, 7–22.the central Montana alkalic province kimberlite–alnoite suite. In: Wilcox, R. E. (1999). The idea of magma mixing: history of a struggleBaker, D. W. & Berg, R. B. (eds) Guidebook of the Central Montana for acceptance. Journal of Geology 107, 421–432.Alkalic Province. Montana Bureau of Mines and Geology Special Publication Wilson, M. (1989). Igneous Petrogenesis. London: Chapman & Hall, 450100, 93–109. pp.
Schultz, C. H. (1962). Petrology of Mt. Washburn, Yellowstone National Wood, B. J. & Banno, S. (1973). Garnet–orthopyroxene and ortho-Park, Wyoming. Ph.D. thesis, Ohio State University, Columbus. pyroxene–clinopyroxene relations in simple and complex systems.
Sisson, T. W. & Grove, T. L. (1993). Experimental investigations of Contributions to Mineralogy and Petrology 42, 109–124.the role of H2O in calc-alkaline differentiation and subduction zone Wooden, J. L. & Mueller, P. A. (1988). Pb, Sr, and Nd isotopicmagmatism. Contributions to Mineralogy and Petrology 113, 143–166. compositions of a suite of Late Archean, igneous rocks, eastern
Smedes, H. W. & Prostka, H. J. (1972). Stratigraphic framework of Beartooth Mountains: implications for crust–mantle evolution. Earththe Absaroka Volcanic Supergroup in the Yellowstone National and Planetary Science Letters 87, 59–72.Park Region. US Geological Survey Professional Paper 729-C, 1–33. Wooden, J. L., Mueller, P. A., Hunt, D. K. & Bowes, D. R. (1982).
Spera, F. J. & Bohrson, W. A. (2001). Energy-constrained open- Geochemistry and Rb–Sr geochronology of Archean rocks from thesystem magmatic processes I: general model and energy-constrained interior of the southeastern Beartooth Mountains, Montana andassimilation and fractional crystallization (EC-AFC) formulation. Wyoming. Montana Bureau of Mines and Geology Special Publication 84,Journal of Petrology 42, 999–1018. 45–56.
Stewart, R. B., Price, R. C. & Smith, I. E. M. (1996). Evolution of York, D. (1969). Least squares fitting of a straight line with correlatederrors. Earth and Planetary Science Letters 5, 320–324.high-K arc magma, Egmont volcano, Taranaki, New Zealand:
702
FEELEY et al. PETROGENESIS OF CALC-ALKALINE MAGMATISM
regressions used for determining the isochron values wereAPPENDIX: ANALYTICAL METHODScalculated following the method of York (1969).40Ar/39Ar analytical procedures
The 40Ar/39Ar analyses were carried out at the Universitede Lausanne. Samples together with the standards were
Whole-rock and mineral geochemistryirradiated for 20 MWH in the central thimble positionanalytical proceduresof the US Geological Survey Triga reactor in Denver,
Colorado (Dalrymple et al., 1981). All analyses were made Compositions of silicate mineral phases in Washburnusing a low blank, double vacuum resistance furnace and rocks are given in Tables 3–6. Mineral compositionsmetal extraction line connected to a MAP 215-50 mass were determined at the Universite de Lausanne, on aspectrometer using an electron multiplier. The samples Cameca SX50 electron microprobe using ZAF on-linewere incrementally heated in the furnace and the gas data reduction and matrix correction procedures. Awas expanded and purified using activated Zr/Ti/Al 15 kV accelerating voltage was used with a 10 nA speci-getters and a metal cold finger maintained at liquid men beam current for plagioclase and 20 nA for all othernitrogen temperatures. Linear time zero regressions were minerals.fitted to data collected from eight scans over the mass Major element data, trace element data, modal data,range 40–36. Peak heights above backgrounds were and Sr and Nd isotopic ratios for Washburn igneouscorrected for mass discrimination, isotopic decay and rocks are given in Table 2. All samples were pulverizedinterfering Ca-, K- and Cl-derived isotopes of Ar. Blanks in a shatterbox, tungsten carbide for XRF and isotopicwere measured at temperature and subtracted from the ratio determinations, and alumina for INAA. Majorsample signal. For mass 40, blank values ranged from and trace analyses by XRF were determined at the4 × 10−15 moles below 1350°C to 9 × 10−15 moles at Geoanalytical Laboratory at Washington State University1650°C. Blank values for masses 36–39 were below 2× following the procedure described by Johnson et al. (1999).10−17 moles for all temperatures. Isotopic production INAA was performed at the Oregon State Universityratios for the Triga reactor were determined from analyses Triga reactor facility following the method of Laul (1979).of irradiated CaF2 and K2SO4 and the following values Estimated precision on the INAA analyses is better thanhave been used in the calculations: 36Ar/37Ar(Ca) = 5% (1� standard deviation) for Sc, Co, Cs, La, Sm, Eu,0·000264 ± 0·0000017; 39Ar/37Ar(Ca) = 0·000673 ± Tb, Yb, Lu, Hf, and Ta, and 7–12% for Ce, Nd, and0·0000037; 40Ar/39Ar(K)= 0·00568± 0·004. Correction U. Sr and Nd isotopic ratios were determined at thefor the neutron flux was determined using the standard Keck Center for Isotope Geochemistry at the UniversityMMHB-1, assuming an age of 520·4 Ma (Samson & of California, Los Angeles, by TIMS, following the pro-Alexander, 1987). A mass discrimination correction of cedure described by Feeley & Davidson (1993). Precision1·008 a.m.u. was determined by online measurement of on the Sr and Nd isotopic determinations is generallyair and was applied to the data. Age plateaux were better than ±0·00002 and ±0·00001, respectively. Allcalculated for samples in which three or more consecutive initial ratios were calculated assuming an age of 53 Ma.steps yielded statistically indistinguishable ages at the Modal data were determined by point counting 1000–95% confidence level and which collectively made up at 1200 points per thin section, with phenocrysts defined
as >0·3 mm in the longest dimension.least 50% of all 39Ar released from the sample. The linear
703