Geochemical evolution of topaz rhyolites from the Thomas

American Mineralogist, Volume 69, pages 223-236, 1984

Geochemical evolution of topaz rhyolites from the Thomas Rangeand Spor Mountain, Utah

ERIC H. CHRISTIANSEN1, JAMES V. BIKUN2, MICHAEL F. SHERIDAN AND DONALD M. BURT

Department of Geology, Arizona State UniversityTempe, Arizona 85287

Abstract

Two sequences of rhyolite lava flows and associated pyroclastic deposits are exposed inthe Thomas Range (6 m.y.) and at Spor Mountain (21 m.y.) in west-central Utah. Bothcontain topaz indicative of their F-enrichment (>0.2%) and aluminous nature. Therhyolites are part of the bimodal sequence of basalt and rhyolite typical of the region.Moderate changes in major elements coupled with large variations in trace elements invitrophyres from the Thomas Range are generally consistent with fractionation of observedphenocrysts. Especially important roles for the trace minerals are suggested. Nonetheless,disagreement between major and trace element models regarding the degree of crystalliza-tion as well as improbably high DTa and DTh suggest that some diffusive differentiationinvolving the migration of trace elements complexed with volatiles also occurred or thatmonazite, titanite or samarskite fractionation was important. Higher concentrations of F inthe evolving Spor Mountain rhyolite drove residual melts to less silicic compositions withhigher Na and Al and promoted extended differentiation yielding rhyolites extremelyenriched in Be, Rb, Cs, U, Th, and other lithophile elements at moderate SiO2 concentra-tions (74%).

Introduction

During the Late Cenozoic much of the western UnitedStates experienced eruptions of bimodal suites of basaltand rhyolite (Christiansen and Lipman, 1972) consequentto the development of the Basin and Range province, theSnake River Plain, and the Rio Grande rift. These rocksare especially characteristic of the Great Basin wherebimodal rhyolitic and mafic volcanism commenced about22 m.y. ago (e.g., Best et al., 1980). Those rhyoliteswhich are rich in fluorine (up to 1.5 wt.%) and lithophileelements (Be, Li, U, Rb, Mo, Sn, and others) appear tobe part of a distinct assemblage of topaz-bearing rhyolitelavas that occurs across much of the western UnitedStates and Mexico. The general characteristics of topazrhyolites have been discussed by Christiansen et al.(1980, 1983a,b). Perhaps the best known are those fromthe area near Topaz Mountain in west-central Utah (Fig.1).

Two age groups of topaz rhyolite lava flows and theirassociated tuffs occur in this region, one forming theThomas Range and the other peripheral to Spor Mountain(Fig. 1). The rhyolites of the Thomas Range have beenmapped as the Topaz Mountain Rhyolite (Lindsey, 1982,

1 Present address: Department of Geology, University ofIowa, Iowa City, Iowa 52242.

2 Present address: Shell Oil Co., P.O. Box 2906, Houston,Texas 77001.

0003-004X/84/0304-0223$02.00 223

1979). The lavas and tuffs were emplaced 6 to 7 m.y. agofrom at least 12 eruptive centers (Lindsey, 1979) andcoalesced to form a partially dissected plateau with anarea in excess of 150 km2. Turley and Nash (1980) haveestimated their volume to be about 50 km3.

The topaz rhyolite of the Spor Mountain Formation hasan isotopic age of 21.3±0.2 m.y. (Lindsey, 1982). Thisrhyolite and its subjacent mineralized tuff (the ''berylliumtuff") occur in isolated and tilted fault blocks on thesouthwest side of Spor Mountain and as remnants ofdomes and lava flows on the east side of the mountain.Lindsey (1982) suggests that at least three vents wereactive during the emplacement of the rhyolite of SporMountain. Eruptive episodes for both the Spor Mountainand Thomas Range sequences commenced with the em-placement of a series of pyroclastic breccia, flow, surge,and air-fall units and were terminated by the effusion ofrhyolite lavas (Bikun, 1980).

This study documents the Petrologic and geochemicalrelationships of these lavas. Particular attention is paid tothe role of fluorine in the evolution of these rocks and tothe origin of their extreme enrichment in incompatibletrace elements. We conclude that most of the variation isthe result of the fractionation of the observed pheno-crysts; "thermogravitational diffusion" appears to haveplayed only a small role.

The petrology of these rhyolites was initially studied aspart of an investigation of uranium deposits associatedwith fluroine-rich volcanic rocks (Burt and Sheridan,

224 CHRISTIANSEN ET AL.: TOPAZ RHYOLITES

LEGEND

Quaternary alluvium

P \ ^ l Topaz Mtn. Rhyolitei-'* -fM (6 m.y.)

• Spor Mtn. Rhyolite(21 m.y.)

i v v Older,volcanic rocks.(30-42 m.y.)

Sedimentary Rocks

Location in Utah

After Lindsey (1979)

Fig. 1. Generalized geologic map of the southern Thomas Range, Utah (after Lindsey, 1979, 1982; Staatz and Carr, 1964). Samplelocalities shown by dots (surface) and crosses (drill core). Sample numbers correspond to those in Tables 1 and 2. Samples 61a, b,and c were collected from a site north of the area shown and sample 6-358 and 18-113 are from a drill core obtained from a site just tothe west of the map area (Morrison, 1980).

1981). The rhyolite of Spor Mountain is underlain by acogenetic tuff which is mineralized with Be, F, and U(Lindsey, 1977, 1982; Bikun, 1980) and is associated withsedimentary uranium deposits and small fluorite deposits.Presumably the lavas or their intrusive equivalents werethe source of the ores. In contrast, the Topaz MountainRhyolite is not associated with economic surface mineral-ization, but is similar in many respects to the older lavas.There is also increasing evidence that such fluorine-richvolcanic rocks are intimately associated with subvolcanicporphyry molybdenum deposits (Burt et al., 1982).

Analytical methods

The samples analyzed in this study were selected froma suite of more than 100 samples collected from outcropsin addition to drill-core samples obtained by Bendix FieldEngineering Corporation (Morrison, 1980). Samples se-lected for chemical analysis were chosen to represent thespectrum of fresh rock compositions on the basis ofPetrographic examination.

Wavelength dispersive XRF was used to determine Si,Ti, Al, Fe, Mn, Ca, K and P. Analyses were performedusing glass discs (Norrish and Hutton, 1969). Energydispersive XRF with undiluted rock powders was used todetermine Rb, Sr, Y, Zr, Nb and Ga. Atomic absorptionspectrometry was used to analyze for Na, Mg, Be, Li andSn. Natural rock powders were used as standards in eachcase. Fluorine was analyzed using ion-selective elec-trodes and a few samples were also analyzed for F and Clusing an ion chromatograph as described by Evans et al.(1981). Estimates of analytical precision are included in

the tables. Other trace elements were analyzed by instru-mental neutron activation at the University of Oregon andare considered to be precise to better than ±5% exceptfor Cr and Eu (±10%).

Mineralogy

The mineralogy of the two sequences of rhyolites aregrossly similar. The Topaz Mountain Rhyolite generallycontains sanidine, quartz, sodic plagioclase ± biotite ±Fe-rich hornblende±titanite. Biotite compositions show in-creasing Fe/(Fe + Mg) with differentiation (Turley andNash, 1980). Two-feldspar and Fe-Ti oxide temperaturescorrelate well and range from 630° to about 850°C (Turleyand Nash, 1980; Bikun, 1980). Augite is found in one lavaflow (SM-61) with a high equilibration temperature. Spes-sartine-almandine garnet occurs in some lavas, generallyas a product of vapor-phase crystallization (Christiansenet al., 1980). Titanomagnetite is present as micropheno-crysts in most samples, but ilmenite is rare.

The rhyolites of the Spor Mountain Formation containsanidine, quartz, sodic plagioclase, and biotite. Biotitesare extremely Fe-rich (Christiansen et al., 1980) andsuggest crystallization near the QFM oxygen buffer. Two-feldspar temperatures cluster around 680° to 690° at 1 bar.Magnetite is more abundant than ilmenite.

Magmatic accessory phases in both rhyolites includeapatite, fluorite, zircon, and allanite. Topaz, sanidine,quartz, Fe-Mn-Ti oxides, cassiterite, garnet, and beryloccur along fractures, in cavities and within the ground-mass of the rhyolites. They are the result of crystalliza-tion from a vapor released from the lavas during coolingand devitrification.

CHRISTIANSEN ET AL.: TOPAZ RHYOLITES 225

Rock chemistry

Major element compositionChemical analyses of vitrophyres and fresh felsites

from both volcanic sequences are presented in Tables 1and 2, along with CIPW normative corundum or diopsideas calculated on a water- and fluorine-free basis (Fe2+/(Fe total) = 0.52; Bikun, 1980). All of the samples havehigh Si, Fe/Mg, and alkalies and low Ti, Mg, Ca, and P,features shared with all topaz rhyolites and typical ofrhyolites from bimodal (mafic-silicic) associations fromaround the world (Ewart, 1979; Christiansen et al ,1983a). As expected, both suites are generally high influorine (greater than 0,2%). Only two vitrophyres, onefrom each sequence of lavas, are peraluminous (corun-dum normative). Devitrification of the vitrophyres mayresult in compositions with normative corundum, forexample compare SM-61a, b, and c or samples SM-34 and35 which were collected from different "facies" of thesame lava flows. In both cases the intensely devitrifiedsamples are corundum normative. This is apparentlyrelated to the loss of an alkali-bearing vapor-phase duringtheir subaerial crystallization (cf. Lipman et al., 1969).

Vitrophyres from the two rhyolites can be distin-guished by their major element geochemistry. In general,the rhyolites from the Thomas Range have greater con-centrations of Si, Mg, and Ti, and less Al and F than the

older rhyolites from the Spor Mountain Formation.Whole-rock compositions are plotted in terms of nor-

mative quartz-albite-orthoclase in Figure 2 for compari-son with experimental work in the same system (withH2O or with H2O-F). Normative femic constituents aregenerally less than 2% of the total, so when contoured interms of An content, this systems is probably a reason-able model for the magmas under consideration. Al-though the vitrophyres from the Spor Mountain Forma-tion show considerable scatter, there is no overlap withthose from the Thomas Range. Both suites lie on thefeldspathic side of the hydrous minima. Only sample SM-35, which displays an anomalously high calcium content,lies in the primary phase field of quartz on this diagram.With decreasing Ca, Mg, Ti, and increasing F, the sam-ples from the Thomas Range plot nearer the fluid-saturat-ed minimum. We interpret this trend (also noted byHildreth, 1977) as a reflection of lower temperatures andthe approach to fluid-saturation of an initially unsaturatedmagma crystallizing at about 1 to 1.5 kbar total pressure.The displacement of the Spor Mountain rhyolites towardthe Ab apex is probably the result of their higher fluorinecontent. Manning (1981) has shown that increasing Fconcentration shifts the fluid-saturated minimum at 1kbar to increasingly albitic compositions (Fig. 2). Apoorly constrained path defined by residual melts in acooling granite-HF-H2O system shows this same trend

Table 1. Whole-rock chemical analyses of rhyolites from Spor Mountain, Utah

* V= vitrophyre, F= felsite, T= tuff» S= spherulitic vitrophyre.** L0I= Loss on ignition at 900°C for four hours.*** Calculated from volatile - free analyses.**** Estimated precision based on replicate analyses. Reported as one standard deviation expressed as percentage of amount reported.n.d.= not determined.

Lithology*Sample No.

SiO2

TiO2

Al AFe2°3MnS 3MgOCaONa2OK.,0P2°5Total

LOI**

RbSrYZrNbGa

Cdi

V31

73.30.0613.21.480.070.050.614.024.990.00

97.78

2.96

970592999050

0.07-

F34

74.80.0613.31.520.040.070.603.764.810.00

98.96

0.50

10001013512013550

0.81-

V35

73.90.0613.11.430.060.081.274.333.650.00

97.88

3.55

10605

135120120n.d.

_0.64

F37

75.30.07

12.71.010.080.070.703.854.860.00

98.64

2.57

6055461207340

_0.37

F47

74.50.05

14.40.810.020.080.142.635.950.00

98.58

1.43

16005138513070

CIPW

3.40- .

F63

weight per

73.70.0413.81.170.05.0.210.453.834.910.00

98.16

0.91

parts per

140020568012065

normative

1.40-

M67

cent

65.40.9414.66.350.090.362.873.154.950.00

98.71

2.30

million

5763451453604530

minerals_

1.80

V70

72.20.0413.91.190.060.060.584.904.980.01

97.90

3.00

***

_1.42

V71

73.70.0613.31.330.050.070.704.394.740.01

98.35

2.98

10003012211011445

_0.66

F6-358

74.00.06

13.51.630.070.230.433.924.880.00

98.72

0.77

860409410510535

1.01-

F18-113

74.40.0613.41.480.060.110.594.014.830.00

98.94

0.57

10301014012013075

0.47

F26-112

73.50.0314.01.190.060.140.633.944.850.00

98.34

0.70

145015778713070

1.11-

132551042220

io-

41066415


Table 2. Whole-rock chemical analyses of rhyolites from the Thomas Range, Utah

LithologySample No.

SiO9T±02

AlAFe2°TMnS 3MgOCaONa 0KO6P2°5TotalLOI

RbSrYZrNbGa

cdi

V42

75.90.0912.91.090.080.090.744.114.690.00

99.993.11

58512431157030

_

0.75

T43

75.90.1612.71.170.060.130.743.455.020.00

99.332.66

37419471405040

_

0.34

See footnotes to Table 1.

F58

75.80.0812.90.990.050.170.994,124.820.00

99.920.79

58031611056045

_ •

1.70

V59

76.60.0712.80.980.060.060.673.984.620.00

99.843.34

5331593 .1106035

0.02

V61a

75.00.2813.11.260.040.190.823,485.510.02

99.702.97

18683351904030

_

0.03

F1-93

75.90.0712.81.610.080.130.393.764.830.00

99.570.33

F8-938

75.80.0812.60.930.070.130.693.924.720.00

98.940.34

S29-124

75.90.0713.01.000.070.180.723.654.710.00

99.302.16

parts per million

6146331157245

4402430905145

4705801054840

CIPW Normative Minerals

0.65 _0.44

0.57

V29-206

76.20.0712.70.990.070.050.723.744.800.00

99.342.84

4332.575

1004730

_

0.07

S61b

74.20.2813.11.250.060.261.063.435.460.02

00.122.80

_

0.15

F61c

74.20.3013.81.350.060.210.853.525.560.01

99.860.17

18579372054035

0.49

V62a

76.10.1412.41.040.060.120.763.355.400.00

99.372.83

37725421305045

_

0.70

F62b

75.40.1412,51.000.050.301.973.585.150.00

100.091.66

37259371104220

_. 2.22

S1-63

75.50.1012.71:020.060.140.464.115.020.00

99.110.16

51511241306130

_0.66

(Kovalenko, 1977b). These considerations suggest that1.5 kbar may be a maximum estimate for the crystalliza-tion pressure.

Trace element compositionThe trace element compositions of these rhyolites

(Tables 1,2, and 3) are especially informative; notable arethe substantial enrichment of Rb, Cs, Nb, Y, U, Th, Ta,and HREE and the depletion of Eu, Sr and Co. Thesefeatures distinguish topaz rhyolites from most other rhyo-litic magma types. The highly fractionated character ofthese rhyolites is demonstrated by their similarity to Li-pegmatites and their associated granites (e.g., Goad andCerny, 1981), and to ongonites (topaz-bearing volcanic

• Thomas Range vitrophyres• Thomas Range (elsttes

O Spor Mountain vitrophyreso Spor Mountain felsites

Fig. 2. Normative composition of rhyolites in terms of quartz(Q), albite (Ab) and orthoclase (Or) compared to experimentallydetermined ternary minima in the hydrous system (contoursfrom Anderson and Cullers, 1978) and in a hydrous fluorine-bearing system (Manning, 1981). Numbers on solid contoursindicate PH2O and beside crosses they indicate weight % F inwater-saturated system at 1 kbar. The dashed line is fromKovalenko (1977) and represents the composition of residualmelts formed by fractional crystallization in a hydrous granitesystem with 1 to 2 wt.% HF at 1 kbar and 800 to 550 °C.

rocks from central Asia, Kovalenko and Kovalenko,1976).

Topaz Mountain Rhyolite. Rhyolites from the ThomasRange display considerable variation in their trace ele-ment composition even though they are all fairly enrichedin lithophile (or "fluorophile") trace elements. Signifi-cantly, vitrophyres with low equilibration temperatures,like SM-29-206 (630°C), are more enriched in lithophileelements than those with high-equilibration temperatures.With increasing evolution (as indexed by uranium con-centration) F, Be, Sn, Li, Rb, Cs, Na, Th, Ta, Nb, Y andthe HREE increase whereas K, Sr, Zr, Hf, Sc, Mg, Ca,Ti, P, and the LREE plus Eu are depleted. REE patternsthus become flatter (lower La/Lu) and Eu anomaliesdeeper (lower Eu/Eu*) with increasing evolution in theTopaz Mountain Rhyolite (Fig. 3). Variation diagrams forsome of these elements are shown in Figure 4. Enrich-ment factors for two rhyolites are shown in Figure 5.

The correlation of these changes with the eruptivehistory is the subject of work in progress, but the coher-ent variation of many trace and major elements across thesuite (plus their mineralogic and temporal similarity)suggests that the rhyolites are part of a comagmatic groupof rocks.

Spor Mountain Formation. The samples analyzed fromthe Spor Mountain Formation are more enriched in U,Th, Rb, Cs, Be, Li, Sn, Ta, Ga, Nb, Y and REE(especially HREE) than the younger rhyolites from theThomas Range. The rhyolites of the Spor MountainFormation have extremely low.K/Rb (30 to 44), highRb/Sr (ca. 150), and low Mg/Li (2 to 5). Semiquantitativeanalyses reported by Lindsey (1982) are compatible withthese results. Thus the composition suggests that theSpor Mountain magmas were highly differentiated. The


Table 3. Trace element composition of topaz rhyolites from Spor Mountain and the Thomas Range, Utah

LiCsBe

CrCoSc

SnHfTa

ThUF

ClLaCe

NdSmEu

TbYbLu

La/Li

La/Ce

SM-61a

202.52

n.d0.42.8

27.92.3

255

1900

73132

638.81.2

1.34.00.53

^ 14-°'N 1 > 4

Eu/Eu* 0.35

SM-62a

407.46

n.dn.d1*8

5.54.5

6015

1600

129

35

0.21

1.29.80.90

-

-

0.06

THOMAS RANGESM-29-206

609.77

2.6n.d1.7

4.95.4

4719

3200

1853.1

317.50.065

1.99.61.7

1.1

0.88

0.02

SM-42

14.1

n.dn.d2.2

5.76.4

5625

5400

2564

314.90.059

7.91.6

1.6

1.05

0.02

SM-41

14.2

1.4n.d2.4

6.06.0

6527

6400

2762

355.00.051

0.798.81.5

1.8

1.12

0.025

SM-31

55

3.60.42.6

6.226

6436

8200

60144

5213.50.15

3.315.62.5

2.4

1.07

0.024

Note: All analyses in ppm; ratios relative to chondrites.** Estimated precision, reported as one standard deviation as percentage ofn.d. = not detected.

SPOR MTN.SM-35 SM-70

805852

3.20.42.7

7.125

6938

11000

168059137

5117.80.044

3.415.72.6

2.3

1.11

0.006

100

82

-

30

4526

12500

1060

-

-

-

-

-

amount present.

SM-71

60

65

-

-

6731

8000

-

-

-

-

-

-

**

10210

1053

15•52

345

523

327

433

-

-

Spor Mountain

Fig. 3. Rare-earth element composition of the rhyolites fromthe Thomas Range and Spor Mountain. Note the enrichment ofHREE and the depletion of LREE with increasing depth of theEu anomaly in samples from the Thomas Range. The scale on theright applies to samples from Spor Mountain and the scale on theleft to samples from the Thomas Range. Concentrations arenormalized to 0,83 times Leedey chondrite (Masuda et al., 1973).

Fig. 4. Logarithmic variation diagrams for topaz rhyolitesfrom the Thomas Range (•) and Spor Mountain (+). The linesrepresent least-squares fits to the_data from the Thomas Range*The bulk partition coefficient (D) for each element is derivedfrom the equation D = 1 - m, where m is the slope of the line(see text).


Fig. 5. Enrichment factors for 30 elements derived bycomparing two samples from the Thomas Range (heavy bars).Elemental ratios derived by comparing early and late parts of theeruption of the Bishop Tuff (Hildreth, 1979) are shown forcomparison (fine bars).

anomalous concentrations of lithophile elements are con-sistent with the suggestion that the Be-U-F ores in theunderlying beryllium tuff were formed as these elementswere released and mobilized by devitrification and leach-ing of the lavas (Bikun, 1980; Burt and Sheridan, 1981).Mineralization is absent from the less evolved rhyolites ofthe Thomas Range.

The REE patterns (Fig. 3) also provide evidence ofextreme differentiation in terms of the large negative Euanomalies (Eu/Eu* = 0.03 to 0.01). In addition thepatterns are nearly flat (low La/LuN = 2.5). Similarpatterns are recognized in other highly evolved rhyoliticmagmas (Hildreth, 1979; Keith, 1980; Bacon et al., 1981;Ewart et al., 1977; and others), and are typical of manytopaz rhyolites (Christiansen et al., 1983a).

Petrologic evolution

The origin of the elemental variations and substantialenrichments of incompatible trace elements (those whichhave bulk partition coefficients D less than 1) in topazrhyolites could result from: (1) small and variable degreesof partial melting, (2) crystal fractionation, or (3) liquidfractionation (e.g., thermogravitational diffusion, liquidimmiscibility, or vapor-phase transport). In this sectionevidence is presented supporting the suggestion that thechemical evolution of the Topaz Mountain Rhyolite (andby analogy the rhyolite of Spor Mountain as well) wasgoverned by crystal fractionation possibly acting in con-cert with some form of liquid differentiation.

Topaz Mountain RhyolitePartial melting. Variable degrees of partial melting of a

uniform source could possibly produce the variationobserved in the incompatible trace elements withoutsubstantially altering the major element chemistry of theliquid (as long as no phase was eliminated from the

source). The simplest discriminant between partial melt-ing and fractional crystallization is the behavior of com-patible trace elements (elements with bulk partition coef-ficients D greater than 1). In feldspar-rich rocks, Eu, Ba,and Sr all have D greater than 1. Large ranges in theconcentrations of these elements cannot be produced bypartial melting processes as the range is limited by 1/D(Shaw, 1970; Hanson, 1978).

Figure 6 shows the behavior of Eu relative to U (with avery low D). In quantitative formulations, the fraction ofliquid produced by partial melting or the residual liquidfrom crystallization (both symbolized by /) are inverselyrelated to uranium concentration, if uranium behaves asan incompatible element. As noted below, DLl is slightlygreater than 0; nonetheless the trend of the data is clearlyinconsistent with batch partial melting, but indicates aDEu of about 3 for some type of fractionation process.

There appears to be little evidence of partial meltingprocesses remaining in these highly evolved rhyolites. Adetermination of the partial melting history would requirea detailed examination of a number of less silicic consan-guineous rocks (like SM-61a) or perhaps an examinationof their intrusive equivalents.

Fractional crystallization. Many of the elemental varia-tions such as increasing Na, U, Th, and Rb, and decreas-ing Ca, Mg, Ti, and Eu with increasing SiC>2 content ofthe rhyolites of the Thomas Range are qualitativelyconsistent with fractional crystallization of observed phe-nocrysts. The major and trace element data combine tosuggest that samples SM-42, -62, and -29-206 could bederived from SM-61 by crystal separation. This processwas modeled using a program written by J. Holloway and

EuEuo

Fig. 6. Behavior of compatible (Eu) versus incompatible (U)elements during batch partial melting and fractional crys-tallization. Curves were calculated from

c 1: (batch partial melting) and — = /(D J)

. Co

(fractional crystallization). The Eu-U variation does not appearto be related to variable degrees of partial melting.


M. Hillier that conforms in most respects with sugges-tions of Reid et al (1973). Weighting is based on l/wt.% ofeach component, normalized so that SiC>2 has a weight of1. The major element compositions of the rocks (exceptP2O5) and phenocrysts were used as input for the pro-gram. Feldspar (Ab/An/Or), magnetite, ilmenite and bio-tite (Fe/Mg) compositions were varied within limits deter-mined by microprobe analyses of phenocrysts (Bikun,1980; Turley and Nash, 1980). Table 4 shows the resultsof this attempt and demonstrates that fractional crystalli-zation of quartz, sanidine, plagioclase, biotite, and Fe-Tioxides (magnetite and ilmenite in a 50:50 mixture) canaccount for most of the chemical variation. Mn and Mgshow the largest (but statistically insignificant) relativeerrors. The SiO2 content of the "cumulates" ranges from72 to 74%. These models suggest that from 40 to 60%crystallization of a magma similar to SM-61 could pro-duce residual liquids similar to the other three samples.The relative proportions of the phenocryst phases are inaccord with those generally observed in the vitrophyreswith sanidine > quartz > oligoclase > biotite > Fe-Tioxides. Although apatite was not included in the fraction-

Table 4. Least-squares approximations of the effect offractional crystallization of rhyolites from the Thomas Range

observedderivative (%)

a. SM-61a to SM-62a

sio2

TiO2

Al 20 3

Fe2°3MnO

MgO

CaO

Na2°K20

Total

76.1

0.14

12.4

1.04

0.06

0.12

0.76

3.35

5.40

99.37

calculatedderivative (%)

76.1

0.13

12.4

1.04

0.05

0.11

0.76

3.35

5.40

99.35

b. SM-61a to SM-29-206

SiO2

TiO,

A12°3Fe2°3MnO

MgO

CaO

Na20

h°Total

76.2

0.07

12.7

0.99

0.07

0.05

0.72

3.74

4.80

99.34

c. SM-61a to SM-42

sio2

TiO2

A12°3Fe2°3MnO

MgO

CaO

Na20

K,

Total

75.9

0.09

12.9

1.09

0.08

0.09

0.74

4.11

4.69

99.69

76.2

0.07

12.7

0.99

0.05

0.07

0.84

3.76

4.81

99.50

75.9

0.07

12.9

1.09

0.06

0.06

0.82

4.12

4.70

99.74

fractionatingphases (wt. %)

sanidine

quartz

plagioclase

biotite

Fe-Ti oxides

Residual liquid

sanidine

quartz

plagioclase

biotite

Fe-Ti oxides

Residual liquid

sanidine

quartz

plagioclase

biotite

Fe-Ti oxides

Residual liquid

16.55

10.00

8.66

1.21

0.47

63.19

22.29

11.96

6.06

1.28

0.59

57.73

29,53

18.69

9.21

1.73

0.58

39.95

ating mineral assemblage (because of constraints imposedby the program), mass balance relations suggest thatremoval of 0.0006 to 0.0012 wt. fraction apatite couldaccount for the excess P2O5 in the calculated composi-tions. Although these models are in no way unique, theydo suggest that substantial fractional crystallization couldhave occurred without drastically changing the majorelement (Si, Al, K, Na) composition of the derivatives—aqualitative argument usually proffered to discount frac-tional crystallization of granitic magmas to produce sub-stantial enrichments of incompatible elements. AlthoughHildreth (1979) has made a strong case for the lack ofcrystal settling in the magma chamber of the Bishop Tuff,there is considerable evidence (field, theoretical, andexperimental) that many granitic intrusions solidify in-wards by crystallization on the walls and especially thefloor—without significant crystal settling (Bateman andChappell, 1979; Bateman and Nokleberg, 1978; McCarthyand Fripp, 1980; Groves and McCarthy, 1978; Pitcher,1979; McBirney, 1980). Theoretically, the magma cham-ber that gave rise to these topaz rhyolites could havecrystallized substantially between eruptions of the lavas.In fact, the elevated concentrations of fluorine wouldpromote extensive crystal fractionation. Wyllie (1979) hassuggested that the addition of fluorine to a granitic magmaresults in a substantial reduction of the solidus tempera-ture but produces a smaller reduction of the liquidustemperature. This would extend the duration of fractionalcrystallization of a cooling pluton over a wider tempera-ture (and time) interval. The accumulation of fluorine inthe residual melt would further enhance this effect. Inaddition, fluorine may reduce the viscosity of silicatemelts and increase diffusion rates (Bailey, 1977) whichwould aid crystallization and fractionation.

A more rigorous test (at least one with fewer assump-tions) of the crystal fractionation model can be made byusing the trace element compositions and the analyticalmethod of Allegre et al. (1977). Using the conventionalRayleigh fractionation law and trace element variationdiagrams that use elements with very low bulk partitioncoefficients as "monitors" of fractionation, U and Cs inthis case, the bulk partition coefficient for a fractionatingsystem can be derived without further assumptions asoutlined below. This bulk partition coefficient can then beused to determine / (the degree of fractional crystalliza-tion) and to place some limits on the composition of thefractionating mineral assemblage.

During perfect fractional crystallization the Rayleighfractionation law (Neumann et al., 1954) can be used todescribe the trace element concentration of a differentiat-ed liquid, CL, relative to the concentration in the parentmelt, Co:

CL = (1)

where D is the bulk partition coefficient. If D <C 1(incompatible in the mineral assemblage, denoted by *),

230 CHRISTIANSEN ET AL.; TOPAZ RHYOLITES

then:

(2)

Thus, as noted earlier the concentration of a hygromag-matophile element serves as an index of the degree ofdifferentiation. Replacing equation (2) in equation (1)eliminates / and expresses the Rayleigh law in terms ofthe relative concentration of two elements:

(3)

(4)

This can be expressed in logarithmic form as:

In CL = In Co + (1 - D) In etCJ

This is a linear equation in In CL - In CL, if D remainsconstant, throughout the process of fractionation. Theslope of the line is given by (1 - D). Theoretically, thisallows the determination^ D for any element. Examplesof diagrams like these using samples from the ThomasRange are shown in Figure 4. A linear regression tech-nique was used to determine the correlation and the slopeof each variation line; the results are shown in Table 5.The elements Mg, Ca, K, Na, Sc, and Hf have correlationcoefficients less than 0.85. Although such correlationsmay be statistically significant, they were not deemeduseable.

The most critical factor that may affect the applicabilityof these derived D's is the assumption that Du = 0. Usingthe average modal mineralogy and mineral partition coef-ficients from Hildreth (1977) and Crecraft et al, (1981), theprobable D u in the lavas ranges from 0.06 to 0.16(including up to 0.5% allanite or 0.04% zircon). Althoughthese deviations may result in relative errors of about 0 to

Table 5. Derived bulk distribution coefficients (D) for rhyolitesof the Thomas Range

20% in the range of /-values considered here (Allegre etal,, 1977), they do not substantially affect the conclu-sions. In addition, the presumably low DCs and thecorrelation of Cs and U tend to support the assumption ofsmall D's for both elements.

Using the D's of Table 5 and the element concentra-tions of Tables 1 through 3, / (the fraction of liquidremaining in a crystallizing system) can be calculatedfrom equation (1). The results of these calculations,which use SM-61a as a parent, are shown in Table 6 andare compared with / ' s derived from the major-elementmixing model. The calculated / ' s for each sample have atotal range of about 0.1, except when using the data for F,La, and Th (in one instance). Although both the majorand trace element models suggest that the rocks becomemore evolved in the sequence SM-6 la, -62, -29-206, -42,there are large discrepancies between the /-values pre-dicted by the two different types of calculations. Evenincluding the possible 10 to 20% uncertainty produced byassuming the D for uranium equals 0, the two solutionsappear exclusive. In fact to satisfy the / predicted by themajor element model the D for U would have to be lowerthan 0 (an impossible situation), not higher as it may be inreality.__ A series of equations may be defined from the derivedD's and the observed phenocryst assemblage. Theseequations take the form

D1 = Xh +

where D^P = the partition coefficient of element i be-tween phase a and liquid P, Xa = weight fraction of phasea. A set of these equations for various elements may besolved simultaneously to place limits on the compositionof the fractionating mineral assemblage. We have at-tempted to do this using the mineral/liquid partitioncoefficients for high silica rhyolites of (Hildreth, 1977;Crecraft et al., 1981; Mahood and Hildreth, 1983). Solu-tions of these equations are generally consistent with therelative proportions of the major fractionating phases as

Table 6. Comparison of /-values (residual liquid) derivedfrom trace and major element models of differentiation

Note: Calculation of D explained in text; r is a correlationcoefficient obtained during linear regression of data.

Element

uCsBeLiLuTaRbFThMnLaTiEu

Mean _f**

f^-value relative

D

0-0,060.110.240.270.380.360.360.500.631.451.842.94

Major element model f

SM-6 2

0.320.340.290.400.480.340.331.42*0.17*0.340.10*0.440.41

0.37 ±0.63

* Not included in computing mean.** Mean ± 1 standard deviation.

to SM-61

SM-29-206

0.260.260.240.240.200.250.270.48*0.280.220.04*0.190.21

0.06 0.24 ± 0.030.58

SM-42

0.200.18-_

0.220.200.180.16*0.200.150.11*0.260.21

0.20 ± 0.030.40


determined from the major-element modeling (X sanidine-0.46, plagioclase -0.20, X quartz -0.30, X biotite-0.03, and X Fe-Ti oxides -0.01). This mineral assem-blage gives acceptable D's for Be, Cs, Rb, Sr, Mn, Li, Ti,Zr, Eu, and K. The La variation limits the maximumproportion of allanite (which effectively includes anymonazite as well) to 0.0004 (X all = D/D*ifanite). Themaximum proportion of fractionating zircon is limited toabout 0.0004 by the Lu variation. The zirconium deple-tion of about 90 ppm implies a similar proportion(0.00043) of zircon (assuming / —0.25 for the change).This amount would result in early cumulates with Zrconcentrations of about 210 ppm and thus seems reason-able. The P depletion suggests that apatite made up about0.00065 weight fraction of the cumulate (implying about0.03% P2O5 in the cumulate).

Bulk distribution coefficients for Y, Nb, Ta, and Thcalculated using this assemblage of minerals are signifi-cantly smaller than the D's derived from the trace ele-ment variation. This observation may simply imply thatthe individual mineral D's employed in the calculationsare unrealistically low. It is not easy to explain DTa of IF*1

in this fashion as they imply X zircon = 0.005 (or DzfrCon= 860) and X allanite = 0.001 (or D^Lite = 1100). Thesevalues are 3 to 5 times larger than those implied by theREE variation noted above. Extremely small quantitiesof fractionating monazite (Mittlefehldt and Miller, 1983),titanite (identified by Turley and Nash, 1980, in someThomas Range rhyolites), xenotime or even samarskite(reported from plutonic equivalents of topaz rhyolites inthe Sheeprock Mountains of Utah, W. R. Griffitts, 1981,written communication) could account for the high D'scalculated for Th, Y, Nb and Ta. Thus although we havenot yet identified monazite or some of the other phasescited here, they have been reported from similar highlyevolved magmas. The small quantities required would beextremely difficult to detect by Petrographic examination.Alternatively, it is possible that some of the trace elementvariation is not produced by crystal/liquid fractionation—some liquid-state process may be operating. Thus thederived D's may be a measure of the "mobility" ofdifferent elements in the magma. Their variation couldreflect their diffusion rates across gradients of P, T, andcomposition, their partitioning into an accumulating fluid,or their ability to form complexes with other migratingelements.

The lack of agreement between the / ' s derived from themajor element and trace element models and the improba-bly high D's derived for some elements may indicatesome sort of differentiation process in the liquid state.However, none of these observations suggest that crystalfractionation did not operate. Indeed the overall majorand trace element systematics imply that any liquiddifferentiation was probably minor and may have onlyaffected some trace elements. Keith (1982) and Nash andCrecraft (1981) came to similar conclusions regarding thedifferentiation of other rhyolitic rocks.

Liquid fractionation. The major processes of liquiddifferentiation that do not depend solely on crystal-liquidequilibria are (1) liquid immiscibility, (2) separation of avapor or fluid, and (3) diffusion (thermogravitationaldiffusion, double-diffusive crystallization, the Soret ef-fect).

The evolution of immiscible silicate and fluoride liquidsis possible in fluorine-rich magmas (Koster van Groosand Wyllie, 1968; Kogarko and Ryabchikov, 1969; Hardsand Freestone, 1978; Kovalenko, 1977b) and could theo-retically play an important role in the trace elementevolution of magmas. However, fluorine concentrationsin hydrous alumino-silicate melts must be extremely highbefore immiscibility occurs (6 wt.% for granite-NaF-H2O, Hards and Freestone, 1978; 3 to 4 wt.% for granite-HF-H2O, Kovalenko, 1977b; 2.5 to 4 wt.% for ongonite-HF-H2O, Kovalenko and Kovalenko, 1976). Such highlevels of fluorine concentration are not reached in theTopaz Mountain Rhyolite (or the more fluorine-enrichedrhyolites from the Spor Mountain Formation) and itappears that silicate liquid immiscibility played a negligi-ble role in their evolution.

The separation of a fluorine-rich fluid and its conse-quent rise to the apical portion of a magma chamber couldhave had an effect on the evolution of topaz rhyolitemagmas. It is difficult to assess the saturation concentra-tions of various F-species in igneous melts in the absenceof direct experimental studies in H2O-free systems.Based on the correlation of SiO2 and F in ongonitic dikerocks, Kovalenko (1973) suggested that fluid-saturationoccurs at 0.5 to 3 wt.% F, increasing with decreasing SiO2

over the range of 75 to 69 wt.%. However, in view of thepreference of F for melts over fluids (Hards, 1976) itseems more likely that H2O would induce the saturationof a fluid which would contain some small amount of F. Itthus seems appropriate to examine the H2O-F-granite (oralbite) system. Wyllie (1979) reports that F increases thesolubility of H2O in albite melts and Manning (1981) hassuggested an imprecise lower limit for fluid saturation of 4wt.% H2O in H2O-F-granite system. The late crystalliza-tion of biotite relative to quartz and feldspars and theexperiments of Maaloe and Wyllie (1975) and Clemensand Wall (1981) in F-free systems suggest that H2Oconcentration in these granitic melts may have been lessthan this. The position of the samples analyzed here in theQ-Ab-Or system is also consistent with the suggestionthat the magma was not fluid-saturated. The emplace-ment of the rhyolites as lava flows instead of as pyroclas-tic deposits is also suggestive of a water-poor condition(Eichelberger and Westrich, 1981) and the absence of afree-fluid prior to venting. Hildreth (1981) has reviewedother evidence indicating that some magma chamberswhich give rise to large ash flows were not fluid-saturat-ed. Nonetheless, Holloway (1976) has shown that evensmall quantities of CO2 in granitic magmas will producefluid-saturation. In the light of our ignorance of magmaticCO2 concentrations it is worth pointing out that the


accumulation of a fluid near the roof of a magma chamberwould result in net dilution of the magmatic concentra-tions of all trace elements except in the case of Dm/fl(partition coefficient between melt and fluid) less than 1.Kovalenko (1977a) has shown experimentally that Li andRb have fairly high Dm/fl in an ongonite-H2O-HF systemat 1 kbar. With temperature decreasing from 800° to600°C, Dj^fl decreases from 50 to 5 and D£/k decreasesfrom 250 to 50. Thus it seems impossible for Dm/fl to attainvalues of less than 1 before a melt crystallized complete-ly. Webster and Holloway (1980), Flynn and Burnham(1978) and Wendlandt and Harrison (1979) have shownthat high Dm/fl are common for REE in H2O-C1, H2O-F,and CO2-rich fluids at crustal pressures. On the otherhand, Candella and Holland (1981) and Manning (1981b)have shown that Cu, Mo, and Sn prefer coexistinghydrous fluids (±F, Cl) over silicate melts. Thus atransient vapor-phase preceding or coincident with erup-tion may have altered the concentrations of these ele-ments but probably did little to effect the concentrationsof other elements including the REE.

If element enrichment/depletion patterns can be con-sidered indicative (and there is no evidence that it is sosimple), the apparent similarity of the variations in therhyolites from the Thomas Range to those of the BishopTuff (Hildreth, 1979) could be taken as evidence that"convection-aided thermogravitational diffusion" pro-duced some of the characteristics of the topaz rhyolites(Fig. 5). Samples SM-41 or SM-29-206 could representeruptions from the upper part of a magma chamber zonedat depth to a composition similar to SM-61a. However,the uncertainty in the temporal relationship among thesampled lava flows as well as the rapid reestablishment ofchemical zonation in some magma chambers followingtheir eruption (Smith, 1979), precludes postulating thatthe lavas were erupted from a unitary system like theBishop Tuff magma chamber. The samples may simplyrecord temporal variations in a magma chamber like thosedescribed by Mahood (1981) and Smith (1979). Whetherthe process which gave rise to these features (Fig. 5) wasthermogravitational diffusion (Shaw et al., 1976; Hildreth,1979) or crystallization in a double-diffusive system withthe rise of evolved less-dense liquid from the walls of acrystallizing magma chamber (Chen and Turner, 1980;McBirney, 1980; Nash and Crecraft, 1981; Christiansen,1983), or some other process is difficult to evaluate untilmore quantitative formulations of these hypotheses canbe made.

In any case, fluorine probably played an important rolein the differentiation process because of its tendency toform stable complexes with a number of elements, includ-ing the REE (Smirnov, 1973; Mineyev et al., 1966;Bandurkin, 1961). Halogen complexing (and subsequenttransport in a vapor-phase, or within a melt, downthermal, chemical, or density gradients) has been invokedto explain the incompatible trace-element enrichment ofevolved magmas and pegmatites (Bailey and Macdonald,

1975; Fryer and Edgar, 1977; Taylor et al., 1981; contrastwith Muecke and Clarke, 1981) and the depleted nature ofsome residual granulites (Collerson and Fryer, 1978). Forthe REE, the stability of the fluoride complexes increasesfrom La to Lu and at lower temperatures (Moller et al.,1980). Mineyev et al. (1966) experimentally demonstratedthe increased mobility of Y (as representative of HREE)relative to LREE (La and Ce) in acidic solutions with highfluorine and sodium activities across a temperature gradi-ent. Thus, the pronounced HREE enrichment (consid-ered to be a fingerprint of liquid fractionation in rhyoliticmagmas) during the evolution of the Topaz MountainRhyolite may have been the result of the co-migration ofF with HREE (and presumably with Li, Be, U, Rb, Sn,etc.) to the apical portions of a magma chamber. Fromstudies of rhyolitic ash-flow tuffs, Mahood (1981) hassuggested that the dominant control on ash-flow trace-element enrichment patterns may be their roofward mi-gration as volatile complexes. The most important vola-tiles in highly evolved granitic magmas are probably H2O,HF, HC1, B, and CO2(?). The suite of elements associatedwith these volatiles should differ depending on the natureof the complexes or ion-pairs formed by each volatile(Hildreth, 1981).

In another paper (Christiansen et al., 1983b), we sug-gest that the trace element evolution of rhyolitic magmasis sensitive to F/Cl in the melt. With differentiation, F-dominated systems (F/Cl greater than 3) display increasesin Al, Na, Li, Rb, Cs, Ta, Th, and Be, while Cl-dominated systems are usually peralkaline and are associ-ated with greater enrichments of LREE, Na, Fe, Ti, Mn,Zn, Nb, and Zr. Although magmas with variable propor-tions of H2O, F, Cl and CO2 certainly exist, topazrhyolites appear to represent the fluorine-dominated endmember. However, it is important to note that the vola-tiles and alkalies also affect the stability of variousmineral phases important to the chemical evolution ofthese rhyolites (allanite, zircon, biotite, titanite, apatite,pyroxene, and hornblende). Thus, it is not an easy task todistinguish liquid-state processes from those that resultfrom fractional crystallization of a distinctive mineralassemblage which was stabilized by different volatile-element concentrations or ratios.

We suggest that the extreme compositions observed inlavas from the Topaz Mountain Rhyolite are the result ofboth extensive fractional crystallization and at least tosome extent the result of the diffusive transfer of traceelements. Although some of the magmas may have beenfluid-saturated shortly before eruption, the role of aseparate fluid phase in their chemical evolution appearsto have been minor.

Spor Mountain FormationThe limited number of trace element analyses for

samples of the Spor Mountain Rhyolite preclude detailedpetrochemical modeling of its differentiation. However,comparisons with the younger Topaz Mountain Rhyolite


and with Asian "ongonites" suggest some ideas about itsevolution.

The composition of the Spor Mountain Rhyolites im-plies that they evolved from magmas slightly differentfrom those erupted in the Thomas Range, They are richerin Fe, Cr, Co, Ba, LREE, and depleted in Si relative tothe evolved Thomas Range lavas, trends which are incon-sistent with the continued differentiation of the TopazMountain Rhyolite. In contrast, they are extremely en-riched in U, Rb, Th, Cs, Li, Be, Sn, Ta, and the HREE.The trace element compositions do not lie on the trendsdefined by the Topaz Mountain Rhyolite, implyingthat D's were different, that initial concentrations variedor both. However, the overall chemical and mineralogicsimilarity of the two sequences (REE patterns, enrich-ment and depletion of the same elements relative to otherrhyolite types, Fe-enriched mafic minerals) suggests thatthe Spor Mountain magma and the rhyolites of theThomas Range may have evolved from grossly similarparents. The most important features of the older rhyo-lites that require explanation are (1) the low Si and high Alcontents, (2) Na/K greater than 1, and (3) the extremeenrichment of "fluorophile'' elements.

The high Na/K (relative to other topaz rhyolites) of theSpor Mountain Rhyolites can be explained in terms of theeffect of fluorine on residual melt compositions in thegranite system (Manning 1981; Kovalenko 1977b). Bothsets of experiments show a regular increase in the size ofthe quartz field in the Q-Ab-Or system and increasinglyalbitic residual melts with increased fluorine content (Fig.2). Thus as the activity of the fluorine increases in aresidual melt, either by differentiation or by the isochoriccrystallization of a fluid-saturated melt (Kovalenko,1973), quartz fractionation becomes more important lead-ing to a reversal in the normal trend of SiC>2 increase withchemical evolution. As a result aluminum increases in theresidual melt and continued fractionation of sanidineleads to higher Na/K in the liquid. This move to moresodic liquid compositions as a result of potassium feld-spar fractionation is petrographically manifest by thecommon mantling of plagioclase by sanidine (anti-rapi-kivi) in the lavas from Spor Mountain.

Each of these trends (decreasing SiO2 and increasingNa and Al with higher F) is observed in the extremely F-rich topaz rhyolites like those at Honeycomb Hills, Utah(Christiansen et al., 1980) and are well developed in F-rich ongonites from central Asia (Kovalenko and Kova-lenko, 1976). These trends are also mimicked during thecrystallization of rare-metal Li-F granites, strengtheningthe suggestion that topaz rhyolites may be associatedwith mineralized plutons at depth.

The Spor Mountain Rhyolite appears to represent theproduct of protracted fractional crystallization under theinfluence of high fluorine activity in the melt. Liquidfractionation is suggested, but not demonstrated, by theflat REE patterns (Fig. 3). Although the magma may nothave been fluid-saturated during the most of its evolution,

the association of numerous tuff-lined breccia pipes(Lindsey, 1982) with the Spor Mountain magmatismsuggests that the magma became saturated during erup-tion as a result of a rapid pressure drop. Enhanced vapor-phase transport of Sn and Mo and perhaps Be, Li, Cs, U,and other elements may have occurred under theseconditions adding to the effect of other differentiationprocesses. This mechanism could lead to the rapid forma-tion of a lithophile-element rich cap to the magma cham-ber. The subsequent eruption of the volatile-rich cap maybe represented by the emplacement of the Be (Li-F-U)-mineralized tuff that underlies the Spor Mountain rhyo-lite.

SummaryThe fluorine-rich rhyolites of the Thomas Range and

Spor Mountain, Utah, are enriched in a distinctive groupof lithophile elements (U, Th, Rb, Li, Be, Cs, Ta and Sn)typical of other topaz rhyolites from the western UnitedStates. The mineralogy and geochemistry of the lavassuggests that they are similar to "ongonites", aluminousbimodal rhyolites, and anorogenic granites.

The topaz rhyolites from the Thomas Range are inter-preted as being derived by differentiation from a lesssilicic but still rhyolitic magma. Calculations suggest thatsanidine, quartz, sodic plagioclase, Fe-Ti oxides, biotite,allanite, zircon and apatite were the dominant fractionat-ing phases over the course of differentiation sampled.However, failure of fractional crystallization models toexplain the high bulk partition coefficients of Ta and Thsuggests that liquid fractionation, aided by the highfluorine content of the magmas may have occurred or thatsome unidentified phase (monazite, titanite, samarskite,xenotime) was removed from the evolving melt.

The Spor Mountain rhyolite contains higher concentra-tions of fluorine and lithophile elements than the TopazMountain Rhyolite. It may have evolved from a magmagenerally similar to the Topaz Mountain Rhyolite, but itsdifferentiation was more protracted. Expansion of thequartz stability field by elevated fluorine concentrationsand the resultant separation of quartz and sanidine mayhave led to its lower Si and higher Na and Al contents.Extended fractional crystallization, possibly aided byliquid fractionation, led to extreme concentrations of Be,Li, U, Rb, Cs and other fluorophile elements. Themagmatic concentrations of these elements set the stagefor the development of the Be-mineralized tuff whichunderlies the lava.

AcknowledgmentsThe authors thank A. Yates, B. Murphy, C. Liu, K. Evans, R.

Satkin, D. Lambert, Z. Peterman and G. Goles for help inperforming the analytical work presented here; D. Lindsey andR. Cole for introducing us to the geology of the Spor Mountainregion; J. Holloway for the use of his computer programs; S.Selkirk for drafting the figures and B. P. Correa, C. Lesher, J. D.Keith, R. T. Wilson, W. Hildreth, W. P. Nash, J. R. Holloway,


and R. T. Gregory for valuable discussions and reviews of themanuscript.

This work was partially supported by U.S. DOE Subcontract#79-270-E from Bendix Field Engineering Corporation andfunds from Arizona State University. The senior author alsoacknowledges support from the National Science Foundation inthe form of a Graduate Fellowship and from the GeologicalSociety of America in the form of a research grant.

References

Allegre, C. H., Treuil, M., Minster, J. F., Minster, B., andAlbarede, F. (1977) Systematic use of trace element in igneousprocess, Part I: Fractional crystallization processes in volcan-ic suites. Contributions to Mineralogy and Petrology, 60, 57-75.

Anderson, J. L. and Cullers, R. L. (1978) Geochemistry andevolution of the Wolf River batholith, a late Precambrianrapakivi massif in north Wisconsin, U.S.A. Precambrian Re-search, 7, 287-324.

Bacon, C. R., Macdonald, R., Smith, R. L., and Baedecker, P.A. (1981) Pleistocene high-silica rhyolites of the Coso volcanicfield, Inyo County, California. Journal of Geophysical Re-search, 86, 10223-10241.

Bailey, J. C. (1977) Fluorine in granitic rocks and melts: Areview. Chemical Geology, 19, 2-42.

Bailey, D. K. and Macdonald, R. (1975) Fluorine and chlorine inperalkaline liquids and the need for magma generation in anopen system. Mineralogical Magazine, 40, 405-414.

Bandurkin, G. A. (1961) On the behavior of rare earth elementsin fluorbearing environments. Geochemistry 1962, 2, 159-167.

Bateman, P. C. and Chappell, B. W. (1979) Crystallization,fractionation, and solidification of the Tuolumne IntrusiveSeries, Yosemite National Park, California. Geological Socie-ty of America Bulletin, 90, pt. 1, 465-482.

Bateman, P. C. and Nokleberg, W. J. (1978) Solidification of theMount Givens Granodiorite, Sierra Nevada, California. Jour-nal of Geology, 86, 563-579.

Best, M. G., McKee, E. H., and Damon, P. E. (1980) Space-time-composition patterns of late Cenozoic mafic volcanism,southwestern Utah and adjoining areas. American Journal ofScience, 280, 1035-1050.

Bikun, J. V. (1980) Fluorine and lithophile element mineraliza-tion at Spor Mountain, Utah. U.S. Department of EnergyOpen-File Report GJBX-225(80), 167-377.

Bodkin, J. B. (1977) Determination of fluorine in silicates byspecific ion electrode following fusion with lithium metabor-ate. Analyst, 102, 409-413.

Burt, D. M. and Sheridan, M. F. (1981) A model for theformation of uranium/lithophile element deposits in fluorine-enriched volcanic rocks. American Association of PetroleumGeologists, Studies in Geology 13, 99-109.

Burt, D. M., Bikun, J. V., and Christiansen, E. H. (1982) Topazrhyolites—Distribution, origin, and significance for explora-tion. Economic Geology, 77, 1818-1836.

Candela, P. A. and Holland, H. D. (1981) The effect of fluorineon partitioning of molybdenum between a magma and ahydrothermal fluid. Geological Society of America Abstractswith Programs, 13, 422.

Chen, C. F. and Turner, J. S. (1980) Crystallization in a double-diffusive system. Journal of Geophysical Research, 85, 2573-2593.

Christiansen, E. H, (1983) The Bishop Tuff revisited: Composi-tional zonation by double diffusive fractional crystallization(DDFC). Geological Society of America Abstracts with Pro-grams, 15, 390.

Christiansen, E. H., Bikun, J. V., and Burt, D. M. (1980)Petrology and geochemistry of topaz rhyolites, western UnitedStates. U.S. Department of Energy Open-File Report GJBX-225(80), 37-122.

Christiansen, E. H., Burt, D. M., and Sheridan, M. F., (1983a)Geology of topaz rhyolites from the western United States.Geological Society of America Special Paper, in press.

Christiansen, E. H., Burt, D. M., Sheridan, M. F., and Wilson,R. T. (1983b). Petrogenesis of topaz rholites. Contributions toMineralogy and Petrology, 83, 16-30.

Christiansen, R. L. and Limpan, P. W. (1972) Cenozoic volca-nism and plate tectonic evolution of the western United States,II. Late Cenozoic. Philosophical Transactions of the RoyalSociety of London, ser. A, 271, 249-284.

Clemens, J. D. and Wall, V. J. (1981) Origin and crystallizationof some peraluminous (S-type) granitic magmas. CanadianMineralogist, 19, 111-131.

Collerson, K. D. and Fryer, B. J. (1978) The role of fluids in theformation and subsequent development of early continentalcrust. Contributions to Mineralogy and Petrology, 67, 151-167.

Crecraft, H. R., Nash, W. P., and Evans, S. H. (1981) LateCenozoic volcanism at Twin Peaks, Utah. Geology and Petrol-ogy. Journal of Geophysical Research, 86, 10303-10320.

Eichelberger, J. L. and Westrich, H. R. (1981) Magmatic vola-tiles in explosive rhyolitic eruptions. Geophysical ResearchLetters, 8, 757-760.

Evans, K. L., Tarter, J. G., and Moore, C. B. (1981). Pyrohydro-lytic-ion chromatographic determination of fluorine, chlorine,and sulfur in geological samples. Analytical Chemistry, 53,925-928.

Ewart, A. (1979) A review of the mineralogy and chemistry ofTertiary-Recent dacitic, latitic, rhyolitic, and related salicvolcanic rocks. In F. Barker, Ed., Trondhjemites, dacites, andrelated rocks, p. 13-121. Amsterdam, Elsevier.

Ewart, A., Oversby, V. M., and Mateen, A. (1977) Petrology andisotope geochemistry of Tertiary lavas from the northern flankof the Tweed Volcano, southeastern Queensland. Journal ofPetrology, 18,73-113.

Flynn, R. T. and Burnham, C. W. (1978) An experimentaldetermination of rare-earth partition coefficients between achloride-containing vapor phase and silicate melts. Geochi-mica et Cosmochimica Acta, 42, 685-701.

Fryer, B. J. and Edgar, A. D. (1977) Significance of rare earthdistributions in coexisting minerals of peralkaline undersatu-rated rocks. Contributions to Mineralogy and Petrology, 61,35-48.

Goad, B. E. and Cerny, P. (1981) Peraluminous pegmatiticgranites and their pegmatite aureoles in the Winnipeg Riverdistrict, southeastern Manitoba. Canandian Mineralogist, 19,177-194.

Groves, D. I. and McCarthy, T. A. (1978) Fractional crystalliza-tion and the origin of tin deposits in granitoids. MineraliumDeposita, 13, 11-26.

Hanson, G. N. (1978) The application of trace elements to thePetrogenesis of igneous rocks of granitic composition. Earthand Planetary Science Letters, 38, 26-43.

Hards, N. J. (1976) Distribution of elements between the fluid

CHRISTIANSEN ET AL.; TOPAZ RHYOLITES 235

phase and silicate melt phase of granites and nepheline syen-ites. N.E.R.C. Report D6, Progress in Experimental Petrology3, 88-90.

Hards, N. J. and Freestone, I. C. (1978) Liquid immiscibility influor-silicate systems. N.E.R.C. Report, Progress in Exper-mental Petrology 4, 11-13.

Hildreth, E. W. (1977) The magma chamber of the Bishop Tuff:Gradients in temperature, pressure and composition. Ph,D.thesis. University of California, Berkeley.

Hildreth, W. (1979) The Bishop Tuff. Evidence for the origin ofcompositional zonation in silicic magma chambers. GeologicalSociety of America Special Paper 180, 43-75.

Hildreth, W. (1981) Gradients in silicic magma chambers. Impli-cations for lithospheric magmatism. Journal Geophysical Re-search, 86, 10153-10192.

Holloway, J. R. (1976) Fluids in the evolution of graniticmagmas. Consequences of finite CO2 solubility. GeologicalSociety of American Bulletin, 87, 1513-1518.

Keith, J. D. (1980) Miocene porphyry intrusions, volcanism, andmineralization, southwestern Utah and eastern Nevada. M.S.thesis. University of Wisconsin, Madison.

Keith, J. D. (1982) Magmatic evolution of the Pine Groveporphyry molybdenum system, southwestern Utah. Ph.D.Thesis. University of Wisconsin, Madison.

Kogarko, L. N. and Ryabchikov, I. D. (1969) Differentiation ofalkali magmas rich in volatiles. Geochemistry International 6,113-1123.

Koster van Groos, A. F. K., and Wyllie, P. J. (1968) Meltingrelationships in the system NaAlSi3O8-NaF-H2O to 4 kbpressure. Journal of Geology, 76, 50-70.

Kovalenko, N. I. (1977a) Measurement of the Li and Rb distribu-tion in the ongonite-H2O-HF system. Geochemistry Interna-tional, 14, 133-140.

Kovalenko, N. I. (1977b) The reactions between granite andaqueous hydrofluoric acid in relation to the origin of fluorine-bearing granites. Geochemistry International, 14, 108-118.

Kovalenko, V. I. (1973) Distribution of fluorine in a topaz-bearing quartz-keratophyre (ongonite) and solubility of fluo-rine in granitic melts. Geochemistry International, 10, 41-49.

Kovalenko, V. I. and Kovalenko, N. I. (1976) Ongonites (topaz-bearing quartz keratophyre)—subvolcanic analogue of rare-metal Li-F granites (in Russian). Nauka Press, Moscow.

Lindsey, D. A. (1977) Epithermal beryllium deposits in water-laid tuff, western Utah. Economic Geology, 72, 219-232.

Lindsey, D. A. (1982) Tertiary volcanic rocks and uranium in theThomas Range and northern Drum Mountains, Juab County,Utah. U.S. Geological Survey Professional Paper 1221.

Lindsey, D. A. (1979) Geologic map and cross-section of Ter-tiary rocks in the Thomas Range and northern Drum Moun-tains, Juab County, Utah. U.S. Geological Survey Miscella-neous Investigation Map 1-1176, scale 1:62,500.

Lipman, P. W., Christiansen, R. L., and Van Alstine, R. E.(1969) Retention of alkalis by calc-alkalic rhyolites duringcrystallization and hydration. American Mineralogist, 54, 286-291.

Maaloe, S. and Wyllie, P. J. (1975) Water content of a graniticmagma deduced from the sequence of crystallization deter-mined experimentally with water-undersaturated conditions.Contributions to Mineralogy and Petrology, 52, 175-191.

Mahood, G. A. (1981) Chemical evolution of a Pleistocenerhyolitic center. Sierra La Primavera, Jalisco, Mexico. Contri-butions to Mineralogy and Petrology, 77, 129-149.

Mahood, G. and Hildreth, W, (1983) Large partition coefficientsfor trace elements in high-silica rhyolites. Geochimica etCosmochimica Acta, 47, 11-30.

Manning, D. A. C. (1981a) The effect of fluorine on liquidusphase relationships in the system Qz-Ab-Or with excess watera lkb: Contributions to Mineralogy and Petrology, 76, 206-215.

Manning, D. A. C. (1981b) The use of radioactive tracers indeterming the behavior of tin in silicate-vapour systems.N.E.R.C. Report D-18 Progress in Experimental Petrology 5,15-16.

Masuda, A., Nakamura, N., and Tanaka, T. (1973) Fine struc-tures of mutually normalized rare-earth patterns of chondrites.Geochimica et Cosmochimica Acta, 37, 239T-248.

McBirney, A. B. (1980) Mixing and unmixing of magmas.Journal of Volcanology and Geothermal Research, 7, 357-371.

McCarthy, T. S. and Fripp, R. E. P. (1980) The crystallizationhistory of a granitic magma, as revealed by trace elementabundances. Journal of Geology, 88, 211-224.

Mineyev, D. A. (1963) Geochemical differentiation of the rareearths. Geochemistry, 12, 1129-1149.

Mineyev. D. A., Dikov. Y. P., Sobolev, B. P., and Borutskaya,V. L. (1966) Differentiation of rare-earth elements undersupercritical conditions. Geochemistry International, 3, 357-359.

Mittlefehldt, D. W. and Miller, C. F. (1983) Geochemistry of theSweetwater Wash pluton, California: Implications for "anom-alous" trace element behavior during differentiation of felsicmagmas. Geochimica et Cosmochimica Acta, 47, 109-124.

Moller, P., Morteani, G., and Schley, F. (1980) Discussion ofREE distribution patterns of carbonatities and alkalic rocks.Lithos, 13, 171-179.

Morrison, B.C. (1980) A summary geologic report on the SporMountain drilling project in Juab County, Utah. U.S. Depart-ment of Energy Open-File Report GJB-19(80).

Muecke, G. K. and Clarke, D. B. (1981) Geochemical evolutionof the South Mountain batholith, Nova Scotia. Rare-earth-element evidence. Cananadian Mineralogist, 19, 133-145.

Nash, W. P. and Crecraft, H. R. (1981) Evolution of silicicmagmas in the upper crust. An experimental analog. EOS, 62,1075.

Neumann, H., Mead, J., and Vitaliano, C. J. (1954) Traceelement variations during fractional crystallization as calculat-ed from the distribution law. Geochimica et CosmochimicaActa, 6, 90-99.

Norrish, K. and Hutton, J. T. (1969) An accurate X-ray spectfo-graphic method for the analysis of a wide range of geologicsamples. Geochimica et Cosmochimica Acta, 33, 431-453.

Pitcher, W. S. (1979) The nature, ascent and emplacement ofgranitic magmas. Journal of the Geological Society of London,136, 627-662.

Reid, M. J., Gancarz, A. J., and Albee, A. L. (1973) Constrainedleast-squares analysis of Petrologic problems with an applica-tion to lunar sample 12040. Earth Planetary Science Letters,17, 433-445.

Shaw, D. M. (1970) Trace element fractionation during anatexis.Geochimica et Cosmochimica Acta, 37, 237-243.

Shaw, H. R., Smith R. L., and Hildreth, W. (1976) Thermogravi-tational mechanisms for chemical variations in zoned magmachambers (abstr.) Geological Society of America Abstractswith Programs, 8, 1102.

Smirnov, M. V. (1973) Electrode Potentials in Molten Chlorides.Nauka Press, Moscow.


'Smith; R. L, (1979) Ash-flow magmatism. Geological Society ofAmerica Special Paper 180, 5-27.

Staatz, M. H. and Carr, W. J. (1964) Geology and mineraldeposits of the Thomas and Dugway Ranges, Juab and TooeleCounties, Utah. U.S. Geological Survey Professional Paper415.

Taylor, R. P., Strong, D. F., and Fryer, B. J. (1981) Volatilecontrol of constrasting trace element distributions in peralka-line granitic and volcanic rocks. Contributions to Mineralogyand Petrology, 77, 267-271.

Turley, C. H. and Nash, W, P. (1980) Petrology of late Tertiaryand Quaternary volcanism in western Juab and Millard Coun-ties, Utah, Utah Geological Mineralogical Survey SpecialStudy 52, 1-33.

Webster, E. A. and Holloway, J. R. (1980) The partitioning ofREE's, Sc, Rb and Cs between a silicic melt and a Cl fluid.EOS, 61, 1152.

Wendlandt, R. F. and Harrison, W. J. (1979) Rare-earth parti-tioning between immiscible carbonate and silicate liquids andCO2 vapor. Results and implications for the formation of light-rare-earth-enriched rocks. Contributions to Mineralogy andPetrology, 69, 409-419.

Wyllie, P. J, (1979) Magmas and volatile components. AmericanMineralogist, 64, 469-500.

Manuscript received, March 3, 1983;accepted for publication, September 27, 1983.

Documents

Geochemical evolution of topaz rhyolites from the Thomas