eruptive pegmatite magma: rhyolite of the honeycomb hillso ...american mineralogist, volume 76,...

American Mineralogist, Volume 76, pages I26I-1278, 1991

Eruptive pegmatite magma: Rhyolite of the Honeycomb Hillso Utah

Rocnn D. CoNcooN*Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, Maryland 21218, U.S.A.

W. P. NlsnDepartment of Geology and Geophysics, University of Utah, Salt Lake City, Utah 84112-l183, U.S.A.

Ansrn-lcr

The Honeycomb Hills rhyolite represents differentiation in a highly evolved magma. Apyroclastic sequence 12.5 m thick and a dome of -0.2 km3 occur in western Utah in aregion populated with several Tertiary topaz rhyolites. Phenocrysts consist of quartz, san-idine, and albite (10-500/o total) in a glassy or fine-grained groundmass. Primary pheno-crysts and megacrysts of topaz and fluorsiderophyllite (- lolo total) and accessory phasesusually associated with rare-element pegmatites occur: fergusonite, ishikawaite, columbite,fluocerite, thorite, monazite, and zircon. Whole-rock composition (SiO, : 73.3o/o, TiO, :0.01, AlrO3 : 14.0, Fe,O, : 0.28, FeO : 0.55, MnO : 0.07, MgO < 0.01, CaO:0.42,NarO : 4.59, &O : 4.44, PrO, < 0.01, F: 0.61, Cl : 0.10, and maximum values Rb: 1960 ppm, Cs : 78,Li: 344, Sn : 33, Be : 80, and Y : 156) is peraluminous, highlyevolved, and comparable to rare element pegmatites. Elevated F contents of up to 2.30/oin glass account for low silica and high alumina contents because the granite minimumshifts toward the Ab apex of the Q-Ab-Or ternary with increasing F. Mineralogy anddistribution of trace elements with order of eruption indicate evacuation of a cool (570-610 'C), chemically stratified magma chamber. Chemical variation within the eruptedvolume can be modeled by Rayleigh fractionation of 7 5o/o of the phenocryst phases. Spatialvariation of some elements, notably Li, Be, B, F, and Cs, may be due to volatile transfer.Enrichment of HrO and F in interstitial melt during crystallization reduced viscosity enoughto allow eruption of the highly crystalline lava of the dome.

IurnooucrroN

The Honeycomb Hills in west-central Utah are theproduct of a single eruptive cycle of unusual lavas. Thelavas are silicic, contain high concentrations of F, arecrystal rich, were erupted at low temperature, and containa suite of unusual accessory minerals. The parent magmawas the result of extensive differentiation and was com-positionally zoned. High silica concentration and crystalcontent together with low temperatures would normallymitigate against eruption. The lavas of the HoneycombHills provide an opportunity to examine these unusualfeatures, to assess their effects on eruption of the lavas,and to place constraints on differentiation mechanisms inhighly evolved silicic magmas. In addition, these lavasare a rare example of an eruptive pegmatite magma, andthey provide specific information on crystal-liquid rela-tions in such magmas that are not obtainable from peg-matites themselves.

Gnor,ocrc sETTTNG

The lavas of the Honeycomb Hills were erupted at 4.7Ma in the eastern Basin and Range province (Lindsey,1977; Turley and Nash, 1980). Figure I is a geologic map

t Present address: U.S. Geological Survey, M.S. 956, Reston,

of the Honeycomb Hills. Previously, the rhyolite bodywas considered to be the remains of two or possibly threelava domes (McAnulty and kvinson,1964; Christiansenet al., 1986). However, flow banding orientations revealthat the Hills are the eroded remnants of a single domethat emanated from a central conduit that can be distin-guished by concentric vertical flow banding in coarselycrystalline rhyolite (Congdon and Nash, 1988). There areseveral other volcanic units in the immediate vicinity,but all are older and unrelated to the genesis ofthe Hon-eycomb Hills magma body. An underlying tuff (Ttrt) issimilar in composition to tuffs erupted from the ThomasRange 50 km to the east and dated at 6.1 Ma (Turley andNash, 1980). Latites and dacites of Oligocene age (Hogg,1972) outcrop adjacent to the Honeycomb Hills and areunderlain by the Kalamazoo Tuff dated at 34 Ma (Hag-strum and Gans, 1989). The Kalamazoo Tuff(Tjt) liesunconformably on the Devonian Guilmette Formation,a thick carbonate sequence that overlies about 5 km oflower Paleozoic sediments (Hintze, 1988).

The rhyolites of the Honeycomb Hills can be conve-niently divided into two units, a pyroclastic sequence andthe overlying lavas that form a dome. The proximal py-roclastic sequence consists of 12.5 m of pumice-bearingair-fall tuffs with rare interbedded surge deposits. Thetephra deposit is capped by a vitrophyric breccia thatViryinia 22092, U.S.A.

0003-{04x/9 l /0708-l 26 I $02.00 t 2 6 l

r262

represents a vitric bomb bed marking the end of the ex-plosive phase of the eruption. The upper unit is a flow-banded felsite that forms the single dome; based upon apresent day extent and height (240 m), it had an originalvolume of approximalely 0.2 km3. Any original pumi-ceous carapace has been lost to erosion that has exposedthe internal structure of the dome including the centralconduit.

AN.c.LyrrcAL METHoDS

Whole-rock SiO2, Al2O3, CaO, Fe,o,, TiOr, MnO, Cl,Rb, Y, Zr, and Nb were measured by X-ray fluorescencespectroscopy. MgO and Fe2* were determined volumet-rically, PrO, by colorimetry, NarO and KrO by flamephotometry, Li by atomic absorption, and F by selectiveion electrode. Be, B, As, Cs, Th, U, and rare earth ele-ments (REE) were determined by instrumental neutronactivation analysis (INAA) by Neutron Activation Ser-vices, Toronto, Canada. Electron microprobe analyseswere performed with Applied Research LaboratoriesEMX-SM and Cameca SX-50 instruments at the Uni-versity of Utah, using a combination of natural mineralsand synthetic materials as standards and applying ZAFand Q@z) (PAP) correction procedures. Mineral separatesof biotite, plagioclase, sanidine, topaz, and quartz wereanalyzed for trace elements by INAA. Separation was byuse of heavy liquids and a Franz isodynamic magneticseparator. Separates were examined under a binocularmicroscope, and foreign grains were removed by hand.Small accessory minerals were not always eliminated, asdiscussed below. Modal analyses were made by pointcounting of samples in thin section. The total number ofpoints counted varied between 1000 and 2000, depend-ing on sample size.

Prrnocru.prrv

The proximal stratigraphic section is illustrated in Fig-ure 2, in which the pyroclastic section is shown to scale.Pyroclastic deposits are light colored (white to light gay)when fresh and oxidized to brown or red-brown whenweathered. Pumice fragments up to 50 cm in diameterare abundant, comprising at least 500/o of the deposit.Latite clasts averaging 10 cm in diameter are present inconformable layers up to 50 cm thick, and these oftencomprise up to 500/o of an individual layer. These lithic-rich layers delineate "throat clearing" episodes in the ini-tial stages of the eruptive cycle. Other xenoliths includequartzite and limestone clasts up to 50 cm in diameter.Two distinct varieties of quartzite occur as xenoliths. Thedome contains red, hematite-stained, slightly conglom-eratic quartzite derived from the Prospect MountainQuartzite of lower Cambrian age. Within the tephra arexenoliths of white quartzite from either the EurekaQuartzite (Ordovician) or sandstone lenses in the Guil-mette Formation (Devonian). The Guilmette Formationoutcrops a few kilometers west of the Honeycomb Hills.Based on stratigraphic sections for the adjacent FishSprings and southern Deep Creek Range (Hintze, 1988),

CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA

----\_ contact--------o--l]-T-n.5-1.+ os -.---- approximatecontact

{ adit

Latite and dacite

Kalamazoo tuffCross stratifiedpyroclastic deposits

Fig. 1. Geologic map of the Honeycomb Hills area. Locationis approximately 113'35'W, 39%3'N. For complete key to thepyroclastic section, see Figure 2.

the Eureka and Prospect Mountain Quartzites lie at depthsof about I and 5 km, respectively, below the HoneycombHills. Pumice fragments contain from l0 to 250lo phe-nocrysts consisting of plagioclase, sanidine, quartz, andless abundant F-rich annite and topaz set in a glassygroundmass (Table l). Feldspar phenocrysts ar€ euhedral,whereas quartz is commonly embayed and subhedral.

The dome-forming felsites are markedly flow banded,hght gray in color when fresh, and brown when weath-ered. They are holocrystalline, with phenocryst contentsranging from 30 to 40o/o that consist of sanidine, plagro-clase, quartz, biotite, and topaz in decreasing order ofabundance (Table l). The average grain size of the crys-talline groundmass is 0.3 mm; groundmass topaz occursas needles up to 0.1 mm in length. Xenoliths of quartziteup to 50 cm in diameter tre present in the dome. Cognate(?) xenoliths of pegmatite up to 20 cm in diameter andtopaz-rich xenoliths are present. The pegmatite xenolithscontain miarolitic cavities and are composed of potassi-um feldspar, plagioclase, quartz, biotite, and topaz, typ-ically l-2 cm in diameter. Topaz-rich xenoliths are poi-kilitic in texture and contain up to 200/o topaz togetherwith quartz, potassium feldspar, plagioclase, and fluorite.Megacrysts of topaz, sanidine, and biotite are commonin a zone donoted "Topaz Horizon" in Figure 1. This isa horizon approximately 10-20 m thick and subparallel

lg|l Thhr ll-aala'l

unconsolidated alluvium ** topaz horizonand lake deposits s2. sample site location

Flow banded domal topaz ffi$l Poorly sorted crystalrhyolite flows fi]'et$ll vitric tuff

llflll Air fall pyroclastic1 " " ' ' l depos i t s E

ffi

Fig. 2. Stratigraphic column of the Honeycomb Hills area.Only the tephra section is to scale.

to flow banding. Megacryst sizes are: sanidine, l-3 cmtypical, up to 8 cm; lopaz, up to 3 cm; and biotite, up to2 cm. Vugs and lithophysae, evidence of a discrete vaporphase, are not present in the dome felsite.

TABLE 1. Modal analyses (vol%)

r263

Magnetite is more abundant in the pyroclastic se-quence but is still present only in trace amounts; it iscommonly oxidized. Other trace phases in both units in-clude zircon, thorite, monazite, fluorite, fluocerite, co-lumbite, fergusonite, and ishikawaite.

Mrxnur,ocv

QuartzPhenocrysts of quartz are smokey throughout but ap-

pear to be less often embayed in felsite than in pumice.Neutron activation analyses of quartz separates are givenin Table 2. Quartz contains trace amounts of transitionmetals but in lesser amounts than coexisting sanidine. Uand Th are at the 2 ppm level. Some quartz grains containmelt inclusions that may be responsible for those traceelements as well as the 400 ppm Na reported; Na couldalso be present to balance charge for trace amounts of Al,which is usually present in quartz.

Feldspar

Sanidine and plagioclase (Abrr-n') occur as phenocrystsand in the groundmass. Sanidine phenocrysts are typi-cally l-3 mm in diameter, and plagioclase is usually 0.5-I mm. Feldspars of both varieties are unzoned, euhedral,and without reaction rims.

Chemical analyses are given for feldspar in Table 3.Average mole percent compositions are AnuAbrrOru forplagioclase and Ano rAbrrOru, for sanidine, with little dif-ference between feldspars from pumice or felsite exceptthat the most Ab-rich plagioclase occurs in pumice. AllHoneycomb Hills plagioclase is more sodic (Ab*r-nr) thanplagioclase reported for Macusani glass (Abr.*o) (Picha-vant et al., 1987) or other rhyolite localities in the westernUnited States.

Plagioclase exhibits a negative Eu anomaly (Fig. 3),although Eu in oligoclase was below the limit of detection(0.05 ppm). On chondrite-normalized diagrams, plagio-clase is normally enriched in Eu relative to the other REE(Drake and Weill, 1975; Nash and Crecraft, 1985). Three


.FI

ooo

z

cnN

(6ooo

oz

Samole numbers Lithology description

Unconsolidated fluvialand lake deDosits

.*+S++{ ll:lg:l8+ $??:3i:3?-.,* 33:13:1f

Flow banded domalrhyolite flows

EI!SSITc)4wC)

!

oeu.o@ ug(D^ A L Y - - i 1 F

5 0

4 9

48

47

4 6

4 5

4 443

421

Air fall tuffs. Maficclasts throughout, moreabundant in zones up lo1m wide. Pumice lumpsmore abundant, alsoabundant in zones 1 mwide. Vitrophyre lumpsoccur in great abundanceat the base of the flowrocKs.

PumiceVitrophyre clastsShoshonitic basaltclasts

o@

a{ ? = z I

o < ? Q

*:"f*'25

Cross-stratified HH tuffs

Vitric tuff29 Shoshonitic basalts

Welded, crystal vitric tuff\

Paleozoic carbonatesI \

PumiceVitro-phyre Felsite

Peg-matite

xenolith

1 0631 7

SanidinePlagioclaseQuartzBiotiteTopazFluoriteOpaques

Total

GlassVoidsUndifferentiated

Total

2.343.762.85

tr.tr.Ir.tr.

8.95

59.93 70.10 57.01 59.54 60.7431.12 20.83 22.60 13.49

1 00.0 1 00.0 100.0 100.0 1 00.068.60 65.41 57.90 62.55

100.0 1 00.0 1 00.0 100.0

20.29 16.20 20.68 16.929.43 1 1.30 12.09 6.467.72 9.36 6.52 53.300.76 0.32 0.63 3.18

tr. 0.09 tr. 19.58tr. tr. tr. tr.

0.14 O.32 0.15 tr.38.33 37.59 40.07 99.44

61.67 62.41 59.93100.0 100.0 1 00.0 99.4

2.93 8.79 6.844.13 5.74 9.521 .60 4.95 10.090.41 0.75 0.52

tr. tr. tr.tr . 0.15 tr.tr. tr. tr.

9.07 20.38 26.97

Phenocrysts1 8.90 15.787.22 9.534.62 8.460.66 0.68

tr, tr.tr . 0.15tr. tr.

31.40 34.59Groundmass

13.80 18.34 14.99't2.70 10.12 12.9913.50 7.52 7.991.29 1.46 1.620.66 tr. tr.

tr. tr. tr.0.1 5 tr. 0.15

42j0 37 .45 37.74

15.2110.7112.410.73

tr'tr.

0.2039.26

62.26100.0

Notei Mode for sample 1 0X was calculated by least-squares fit of whole-rock data to analyses of phases present.

1264

TABLE 2. Neutron activation analyses of mineral separates


Pumice Vitrophyre Megacrysts

1 B i 48 Bi 32 Bi 32 Pg 32 Sa 32Qz 32Tz TH Bi TH Tz TH Sa

Sc 120Co 4.7Zn 500A s 4BrRb 4500sb 18 .0Cs 44.0Ba 150Hf 5.2Ta 91w 2 4Au (ppb) 9T h 1 1u 23.1La 14.2Ce 55N d 6Sm 3.55Eu O.2OTb 0.8Yb 9.91Lu 2 .10

134 1144.4 4.0

720 420

?4300 3600

8.4 5.936.9 32.8

150 3401 3 1 892 8710 25< 5 1 425 8230.5 78.031.8 67.3

166 26321 448.53 19.60o.27 0.132.8 6.4

27.1 62.74.82 11.20

0.17 0 .121.1 0 .7

1 3 6< 1 < 1

1200 <101.3 0 .92.0 0.3

20 200.9 0.80.5 <0.5

< 1 < 1< 5 C

2.7 2.22.1 2.1

14.3 2.31 7 63 < 30.49 0.470.11 0 .060.1 0.21 .33 1 .40.29 0.28

0.78 0.151 . 1 1 . 5

17 26< 1 < 11 0 1 0

<10 47001 . 5 1 . 70.5 9.2

<20 50<0.2 <0.2

0.7 <0.5< 1 < 1

290 <5<o.2 <0.2

0.1 0.20.3 0.81 1

<3 <3<0.01 0.02

0.1 0 0 .10<0.1 <0 .1

0.1 <0.050.01 <0.01

0 .161 . 1

1 3a :

300.80.8

<201 . 40.6

< 1<5

o . J

3.334.459I1 .61

<0.050.32 .120.40

5.100.7

28

I1 1 0

1 . 54.9

990272820

b

8568.2

482.0121033660.60 .149.4

92.716.1

1412.6

440201 4

61 008.6

78.2<20

1 . 224031I3.0

34.26.0

< 1<3<0.01<0 .17<o.2

0.410 .13

A/ote-' Samples 1T Bi through 32V Tz are phenocrysts. Samples TH Bi through TH Sa are from single megacrysts. Values in ppm. Bi : biotite, Pg: pfagiocfase, Sa : sanidine, Oz : quartz, fz: topaz.

. Data unusable because of the use of bromoform in seoaration technioues.

factors promote low concentrations of Eu in feldspar atthe Honeycomb Hills: (l) the Honeycomb Hills magmawas at a relatively oxidizing condition (see below), in-creasing the proportion of Eu3+, (2) partition coefrcientsfor Eu decrease with decreasing anorthit€ content in pla-gioclase (Nash and Crecraft, 1985), and (3) the Honey-

TABLE 3. Representative analyses of feldspar

comb Hills magrna has very low concentrations of Eu,often below the detection limit.

Sanidine is the only phenocryst phase without a largenegative Eu anomaly. The Ko for Eu is much higher thanfor adjacent REE (0.48 for Eu, 0.04 for Sm, and 0.04 forTb), although it remains less than l. Reducing the albite

1 0cb6160TH2032484746454342Sample

sio,Alr03FerO"CaoBaONaroK"ORbro

TotalAbOrAn

66.2 66.118.8 18.70.04 0.040.04 0.04

0.033.47 3.42

1 1 . 8 1 1 . 80 .10 0 .11

100.4 100.230.8 30.469.0 69.4o.2 0.2

67.0 67.320.9 20.6

1.50 1.26

10.2 10.50.97 0.89

100.6 1 00.587.4 89.15.5 5.07.1 5.9

577 589

65.7 65.41 8.6 18.5

0.02 0.04

3.44 3.391 1 . 9 1 1 . 90 .15 0 .12

99.8 99.430.4 30.169.5 69.70.1 0.2

67.8 67.019.9 20.4

0.68 1.43

10.7 10.20.93 0.92

100.0 100.091.6 87.95.2 5.23.2 6.9

573 580

65.5 65.81 8.5 18.6

0.040.03 0.05

3.53 3.761 1 .8 11 .40.13 0.09

99.5 99.731.1 33.368.7 66.40.2 0.3

66.8 66.620.4 20.30.04 0.041 .38 1.31

'to.2 10.11 .08 1 .1 1

99.9 99.587.3 87.56.1 6.36.6 6.2

583 621

Alkali feldspar65.7 65.5 65.2 66.118.5 18.5 18.6 1 8.80.03 0.04 0.06o.o2 0.02 0.06 0.05

3.58 3.46 3.52 3.741 1 .8 1 1 .8 11.7 11.40.14 0.1 1 0.07 0.09

99.8 99.4 99.2 100.231.5 30.8 31.3 33.168.4 69.1 68.4 66.70.1 0.1 0.3 0.2

Plagioclase66.9 66.4 66.5 66.920.3 20.2 20.7 20.60.03 0.06 0.051.28 1 .25 1 .61 1 .40

10.3 10.3 10.2 10.11 .08 1 .00 0.95 1 .10

1 00.0 99.2 100.0 100.287.9 88.4 87.0 87.26.1 5.7 5.4 6.16.0 5.9 7.6 6.7

558 567 610 592

65.7 66.1 65.3 65.718.5 18.5 18.6 1 8.6

0.030.06 0.0s 0.04

3.86 3.77 3.76 3.061 1 .3 11.4 1',t.4 12.40.07 0.09 0.10 0.20

99.5 99.9 99.2 100.034.0 33.5 33.4 27.365.7 66.2 66.4 72.70.3 0.3 0.2 0

66.7 66.7 67.1 67.920.8 20.2 20.5 20.1

0.041 .66 1 .25 1.29 0.57

too ro : 103 1 ;0.98 1.02 1.05 0.38

0.031 00.1 99.5 100.2 1 00.186.5 88.3 88.0 95.25.6 s.8 5.9 2.17.9 5.9 6.1 2.7

603 619 584 (524)*

sio,Al2o3FerO"CaOBaONaroK.oRbro

TotalAbOrAnrfc)'

Note.'Dash : concentration < 0.03. End-members in mol7".'Twojeldspar convergence temperature at 1 kbar pressure (Fuhrman and Lindsley, 1988).

-' AB temperature-no convergence.

1 0 0

o

oC

€ro()DtI l

CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA r265

1 000000 1 000000

Thorite t

1 000 1 000Er Yb

REEFig. 3. Chondrite-normalized plots of REE for phenocrysts and accessory minerals. Feldspar, biotite, and quartz data are by

INAA; accessory phases were analyzed by electron microprobe. Eu value for plagioclase is maximum possible (: detection limit).

La EuCe Nd Sm

component ofsanidine apparently reduces the Eu content(Nash and Crecraft, 1985), although the effect is not aspronounced as reducing anorthite in plagioclase.

Mica

The mica of the Honeycomb Hills occurs as pheno-crysts up to I mm in diameter throughout the eruptivesequence and as megacrysts in the topaz horizon of thedome. The mica is unusual in composition because it isextraordinarily enriched both in Fe and F. The high Fcontent of the iron mica attests to the abnormally high-frr/ .frro ratio present in the magma, as discussed below.

Analyses of representative mica samples are presentedin Table 4. Structural formulas are calculated on the basisof 22 O atoms; HrO content is calculated from stoichi-ometry of the OH-halogen site. The micas contain sub-stantial I6lAl and less than six cations in the octahedralsite, resulting in some dioctahedral character. The com-position closely approximates siderophyllite (Deer et al.,1962). Two separates analyzed for Li contained 0.lolo Li.insuftcient to fill the octahedral site. The significant dif-ference between biotite samples in the pyroclastic se-quence and in the dome is that the former contain lessthan 3olo F, whereas the latter average over 5olo F withsome individual grains containing as much as 7.50/o F. Inother respects, the mica samples are of similar composi-tion, with domal biotite containing slightly more Rb andMn and less Fe. A detailed description of HoneycombHills micas is in preparation.

Topaz

Topaz is common throughout the Honeycomb Hillslavas in concentrations up to 0.6 vol0/0. The topaz is nearend-member fluortopaz with 20 wto/o F. In the typicaltopaz rhyolites ofthe eastern Basin and Range province,

Eu Tb LuSm Dy Er Yb

topaz forms by crystallization from the vapor phase incavities in felsite. At the Honeycomb Hills, topaz is aprimary phenocryst phase. Altogether three generationsof topaz are present: megacrysts up to 3 cm in diameter,phenocrysts up to 0.5 mm, and abundant needles up to0.1 mm in the groundmass of the felsite. Most topaz iseuhedral with well-developed terminations and (110) and(210) prisms (Fig. a). Coprecipitation with mica is indi-cated by inclusion of each in phenocrysts of the other.Phenocrysts of topaz also contain inclusions of plagio-clase, sanidine, qlJarlz, zircon, monazite, and columbite,

Fluorides

Primary fluorite occurs as euhedral microphenocrystsof 5-20 pm in diameter in pumice and vitrophyre. Fluo-rite also occurs in anhedral masses up to I mm in di-ameter in the groundmass of some domal felsites and atthe contact between the top ofthe pyroclastic section andthe base of the dome, where it has been concentrated bvhydrothermal processes.

Fluocerite [(Ce,La)Fr] occurs as euhedral micropheno-crysts, 10-50 pm in length in vitrophyre, and as small(<0.1 mm) inclusions in phenocryst biotite in felsite ofthe topaz horizon. Light rare earths and F predominatein this mineral (Table 5). The usual occrrrence of fluoce-rite is in rare element pegmatites (Henderson, 1984), andwe believe this to be the first occrrrence offluocerite inan extrusive rock.

Orthosilicates

Ztcon occurs as 5-50 pm diameter microphenocrystsand as inclusions in mica and niobates. It occasionallycontains micrometer-size inclusions of thorite.

Thorite occurs as discrete grains, 2-10 pmin diameter,associated with zircon and as inclusions from micrometer

La PrCe Nd

La Pr Eu TbCe Nd Sm Dy

Fergusonite.Ce-Niobate o

r266

TABLE 4. Average microprobe analyses of biotite

Pumice Vitrophyre

Sample


- XMg, XSid, XAnn : mole fraction end-members from Munoz (1984)." From Munoz (1984).t From Eouation 1 in text.

to submicrometer size in zircon. Both zircon and thoritecontain substantial IJ, and both preferentially incorpo-rate heavy REE (Fig. 3).

Niobates

Columbite [(Fe,MnXNb,Ta)rOu] occurs as microphe-nocrysts (2-10 pm in length) in vitrophyre, in the ground-mass of felsites, and as inclusions in biotite and topaz; inseveral instances, it occurs in growth contact with ishi-kawaite. Unlike the other coexisting niobates, columbitedoes not incorporate REE or actinides; it does containmore Ta and Mn than coexisting niobates (Table 5).

Fergusonite [an yttrium niobate (Y,C-a,Ce,U,Th)(Nb,Ta,-Ti)O41, ishikawaite [a U-rich variety of samarskite(Y,Ce,U,FeXNb,Ta,Ti)r(O,OH)ul, and an unknown Ce-niobate occur as euhedral, prismatic microphenocrystsup to 50 pm in diameter in vitrophyre and as inclusions

(sample 32, upper photograph) and a biotite inclusion in topaz(sample 62, lower photograph). Upper photograph is approxi-mately 2 mm across; lower is approximately 1.5 mm.

in biotite. These three niobates contain substantialamounts of the heavy rare earths and, collectively, highconcentrations of U, Th, Nb, Ta, and Y (Table 5; Fig. 3).

Phosphate

Monazite occurs as inclusions 5-20 pm in diameter inbiotite and topaz phenocrysts, often in association withzircon and thorite. Monazite contains substantial Th inaddition to light REE.

ColrposrrroN oF LAvAS

Major elements

SiO, contents for rhyolites of the Honeycomb Hillsrange from 70.4 wto/o (72.5o/o, anhydrous) in pumice to76.30/oindomal felsites and from 70.1 to 71.9o/oinglassesnormalized to an anhydrous composition (Table 6). Bycomparison, SiO, contents range from 73.0 to 76.50/o inrhyolite from Spor Mountain, Utah, the closest analoguein composition to the Honeycomb Hitls (Christiansen etal., 1984). On average, silica contents in Honeycomb Hillsare lower than those of less evolved silicic systems in the

sio,Tio,SnO,Alro3FeOMnOMgoCaONaroKrORbroF

HrOSum-o : F,c l

TotalSiratAlSum tet:retAlt l

SnFeMnMgSum oct:NaKCaRbSum int:OHF

Sum hyd:xMg'XSid-XAnn'log[ f",o/f*]--ta"ol

rv(F).'

33.8 36.20.68 0.470.05 0.06

19.0 19 .632.1 26.30 .73 1 .170.14 0 j2

0.040.40 0.779.0 8.70.30 0.592 . 7 3 5 1 30.31 0.182.34 1.27

101 .6 100.5-1.22 -2.20

100.4 98.35.46 5.792.54 2.2'l8 81 .08 1 .490.08 0.060 04.34 3.520.1 0 .160.03 0.035.63 5.260.13 0.241 .85 1.780 00.03 0.062.01 2.082.52 1.351.39 2.60.09 0.054.00 4.000.006 0.0050.85 0.810.15 0 .182.9 2.37

1135 2001.4 0 .90.49 -0.1

34.30.420.06

19.231.90.970.09

0.369.00.372.910.192.31

102.1-1.26

100.8c.5u2.5081 . 1 40.0504.280.130.025.620.111.8400.041.992.471.480.054.000.0040.850.152.86

1 0011 . 40.45

36.1o.770.04

1 8 . 126.41 .900.24

0.438.90.555.61o.270.97

100.2-2.4297.85.862.',t4I1 .310.0903.580.260.06c . J

0.141.8300.062.031.052.880.074.000.0110.740.252.24

1330.8

-0 .16

Fig. 4. Photo micrographs of topaz inclusions in biotite


TABLE 5. Microprobe analyses of representative ac@ssory minerals

1267

A

Zicon

B

Thorite

c

Monazite

D

Fluocerite

E

Columbite

F

Fergusonite

u n

lshikawaite Ceriumniobate

sio,Tio,SnO,ZrO"Hfo,Tho,Uo.FeOMnOCaOPrO,Nb2o5Ta"OuYrO"LarO.CerO.PrrOgNd2o3SmrO3Eu"O.Gd2o3TbrosDYro,ErrOgYb2o3LurO.FSum- o : F

Total

30.1

54.25.490.634.09

0.41

0.040.290.572.000.37

98.2

98.2

18 .4

0 .18

47.424.2o.74

1.480.130.20

0.19o.14

0.38

1.O20.391 . 1 60.520.19

96.50 .1

96.4

3 .13

0.430 . 1 6

12.60.48

0 . 1 025.7

0.851 1 . 1 027.102.289.891.530.383.450.030.73

0.48100.5

0.2100.3

1.250.78

12.207.48

68.19.280.180.100.09

0.06

99.5

99.5

0.79 0.591.52

0.58

0.53

0.9618.8041.40

6.3923.69.472.100.11

39.55 .131.050.030.41

0.730.33

0.500.211.371.272.870.40

97.6

97.6

1.079.27

14.202.230.61

9.551.81

25.799.3

99.3

39.34.971.620.30

12.80

0.370.30

0.96

1.533.262.000.030.36

42.51 .53

23.00.070.700.331.731.200.082.840.666.024.153.560.67

96.9

96.9

2.432.543.450.920.25

96.90 .1

96.8

Note.'A blank value indicates the element was not determined; a dash (-) indicates abundance below the level of detection (approximately 0.03%).A : Zircon inclusion in topaz; topaz horizon felsite TH. B : Thorite microphenocryst; vitrophyre 32. C : Monazite inclusion in biotite phenocryst;vitrophyre 32. D : Fluocerite microphenocryst; vitrophyre 32 Values in o/o element. E : Columbite inclusion in topaz phenocryst; topaz horizon felsiteTH. F = Fergusonite in biotite phenocryst; vitrophyre 32. G : lshikawaite microphenocryst; vitrophyre 32. H : Cerium niobate microphenocryst;vitrophyre 32.

western United States such as the Bishop Tuff (75.5-77.4o/o; Hildreth, 1979); Coso, California (73.8-77.0o/o;Bacon et al., l98l); Thomas Range, Utah (73.3-76.20/o);and Smelter Knolls, Utah (73.3-76.20/o; Turley and Nash,1980). However, lower SiO, contents of highly evolvedmagma are consistent with high F contents, as shownexperimentally by Manning (1981), in which the ternaryminimum in the system quartz-albite-orthoclase shiftedtoward the feldspar join with increasing content of F (Fig.5). We believe that whole-rock F contents do not repre-sent initial magma concentrations or F contents at thetime of eruption. F has been lost from glassy samplesselectively after eruption. Pumices with thin bubble wallscontain negligible amounts of F, whereas massive glassin the vitrophyre contains 2.3o/o F (Table 6). F contentsof melt inclusions in quartz phenocrysts range from 3.0to 3.5o/o in early erupted pumice to I to 30/o in domallavas (Gavigan et al., 1989). These elevated F contentsare similar to concentrations reported for ongonites (Ko-valenko and Kovalenko,1976;' Kovalenko, 1973; Kova-lenko et al., 1971, 1977) and, topaz granites @ichavantand Manning, 1984) and higher than most topaz rhyo-lites. The results presented in Figure 5 are consistent withthe higher F concentrations in the magma volume that

gave rise to the pyroclastic sequence than in the massthat was later erupted passively as domal felsites.

The shift in the ternary minimum toward the feldsparjoin with increasing F results in higher Al and lower silicacontents in early erupted pumice than in later eruptedfelsites. Normative corundum ranges from 0 to 1.50/0. Thetotal alkali content decreases through the eruptive se-quence with a concomitant decline in the Naro/lLrO ratiofrom 1.7 to 1.2. Fe and Mg do not vary significantly, andMg, Ti, and P are present in trace amounts; in all casesthese elements are less abundant than Rb.

The major element composition of Honeycomb Hillslavas is similar to the peraluminous glass from Macusani,Peru (Pichavant et al., 1987; London et al., 1988) exceptthat AlrO, is lower at Honeycomb Hills ( I 3- I 5olo) vs. theMacusani glass (nearly 160lo), which results in higher nor-mative corundum. F contents in vitrophyre are higher atHoneycomb Hills (2.30lo) than in Macusani glass (1.30lo).Despite these general similarities, there are notable ex-ceptions in the extraordinarily high contents of PrOs(0.530/o), BrO3 (0.620/0), and LirO (0.74o/A in Macusaniglass; these elements are l-2 orders of magnitude lessabundant at the Honeycomb Hills. A second rhyolite thatclosely resembles Honeycomb Hills rhyolite in compo-

r268 CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA

sition is at Spor Mountain, Utah, erupted at 2l Ma 50km to the east of the Honeycomb Hills (Christiansen etal., 1984). Webster et al. (1987) studied the phase equi-libria of this composition, and their results are applicableto the Honeycomb Hills system as described below.

Trace elements

The extreme differentiation of the Honeycomb Hillsmagma is shown by enrichments in several elements (Ta-ble 7). Some maximum values (ppm) are: Rb 1960, Cs78,Li344, Sn 33, Be 80, and Y 156. Ti(294 ppm max-imum) and Zr (67 ppm maximum) are examples ofstrongly depleted elements. By comparison, maximumconcentrations of incompatible elements in the BishopTuff are about No those of the Honeycomb Hills: Rb190, Cs 7 .2, Li 37, Sn 3.6, Be 4. 1, and Y 25; concentra-tions of compatible elements are Ti 1260 and Zr 140(Hildreth, 1979). With respect to the Macusani glass (Pi-chavant et al., 1987; London et al., 1988), the lavas ofthe Honeycomb Hills contain more Be, Cl, Sc, Co, Rb,Y, Nb, Sb, Hf, Pb, Th, and REE; the latter are ten to 100times more abundant at the Honeycomb Hills. Macusaniglass is relatively enriched in Zn, As, Sn, Cs, and W.Samples from both localities have similar contents of Zr,Ta, and U and are characterized by very low Sr and Ba.

The systematic variation in elemental abundances withrespect to sequence of eruption at the Honeycomb Hills

TABLE 6. Representative maior element analyses and CIPW norms

AbFig. 5. Quartz-albite-orthoclase ternary with minimum melt

compositions for HrO : I kbar, l0/o F, 2o/o F, and 4o/o F (adaptedfrom Manning, 1981). Filled circles are whole-rock domal fel-site, open circles are pumice, and open squares are electron mi-croprobe analyses of glass.

Or

Vitrophyre

1 G 32G203243

sio,Tio,Al,o3FerO.FeOMnOMgoCaONaroKrOPrOuFclH.O*H,O-Sum-O = F,Cl

TotalSiO, (HrO free):

QuartzCorundumOrthoclaseAlbiteAnorthite

70.70.01

15.10 .160.520.09

<0.010.694.994.86

<0.010.540 . 1 12.570.30

100.640.25

100.3972.5

21.200.35

28.7042.203.42

70.40.01

15.10.040.550.10

<0.011 . 1 05.254.83

<0.010.480 . 1 01.920.40

100.280.22

100.0672.O

19.00

26.9044.403.23

28.400.90

26.2038.802.08

25.50

26.9039.305.27

76.00.01

13.00.530.360.08

<0.01o.444 . 1 14.580.010.850.030 . 1 80.02

100.200.36

99.84

33.500.55

27.1034.802 . 1 2

76.10 0 4

13.00.930.200.06

<0.010.363.804.920.010.560 . 1 1

<0.010.05

1 00 .14o.26

99.88

34.200.83

29.1032.201.72

74.90.03

16.90.820.130.07

<0.011.051 . 1 32.410.022.00

n.d.1.320.02

100.800.84

99.96

56.9010.6014.209.565.08

73.3 72.80.01 0.01

14.0 14 .s0.28 0.530.55 0.310.07 0.07

<0.01 <0.010.42 1.594.59 4.644.44 4.55

<0.01 0.040.61 0.310.10 0.050.80 0.04o.24 0.03

99.41 99.470.28 0.14

99.13 99.3374.7

66.6 66.4<0.01 <0.011 5.4 15.50.19 0.050.34 0.490.08 0.11

<0.01 <0.010.35 0.324.87 4.734.15 4.310.18 0 .170.69 2.360.15 0 .15n.d. n.d.n.d. n.d.

93.0 94.70.32 1.02

92.6 93.671.9 70.9

21.3 21.32.25 2.5

24.5 25.541.20 40.001.74 1 .59

CIPW normative minerals

Note.' Samples 1T through 10X were determined by X-ray fluorescence, flame photometry (for Na"O and GO), select ion electrode (F), sodiumtungstate fusion (H,O"), and drying overnight at 110'C (H.O ). Samples 1G through 32G are microprobe analyses of glasses in pumices and thevitrophyre (sample 32V). Silica has been shown with analyses normalized to 100% for reference and comparison with whole-rock data. F is ignored inthe CIPW calculation; if included, Ca is allocated to CaF., resulting in no normative An, which is an inappropriate representation of two feldspar rhyolites.Analyses of other samples are availabb from the authors.


TABLE 7. Representative trace element analyses of whole-rock samples

1269

66TH2032t u 32G

LiBeB

ScTiMnCoZnAsBrRbSr

ZrNbSnsbCsBaHfTaWAu (ppb)ThULaCeNdSmEuTb

LU

n.d.n.o.n.o.n.o.n.o.n.o.5.7

78.4n.d.

R O

44341 51 929.928.8941 85.700.211 . 6

19 .63.60

30679

1901 060

6.7736

7331 . 1

59Z J

1 . 21 560

21392781251 5

75.9584.3

3522<51 831.823.795za5.560.081 . 3

16.23.03

1 7296.6

1 96037512375335.4

t c . I

3.7

1 91 519.016.2571 33.080.240.79.961.94

n.o.n.o.n.o.n.d.n.o.n.d.7.7

61.4n.d.

b . 5

32za<52226.634.6

139281 1 . 4 00.232.8

31.55.84

34468

1301 030

4.6068

571

251 48.4

1370I

84477422J . 5

39.6

i'JJ

2426.140.2

1393810.80<0.05

2.629.15.22

1 81 33.8

146033854286232.O

34.2125

5.0

62516.238.8

1373210.400.162.8

30.65.30

2421 132

3163.90

5b

637

265

1 5'14401 19148841 44.9

37.O

: '

2418.147.7

137351 1 .400.193.1

31.75.64

1 1 6

281 060

3.88220494

2954.2

1 0501 4

1566785I1 . 4

16.4

6.2

4215.452.7

1554414.300.133.7

30.45.03

n.o.n.o.

260n.d6 .13

n.d.n.d.1 . 6

3931

2943692

5374.38

49545

n.d.oc

190n.d.5.28

n.o.n.o.1 . 5

3122

192267220991

4.84JZ

784

/Vote.'Values in ppm. Li, Be, and Sn by atomic absorption. Cl, Ti, Mn, Rb, Sr, Y, Zr, Nb, and Ba by X-ray fluorescence. Sc, Co, Zn, As, Br, Sb, Cs,Hf, Ta, W, Au, Th, U, and the REE by instrumental neutron activation analyses. Dash : Not detected; n.d. : not determined.t Data unusable because of the use of tungsten carbide or bromoform in preparation. Analyses of other samples available from the authors.

is illustrated in Figure 6. Several signifrcant features areevident in the data presented in Figure 6. Collectively,the data illustrate that the eruption tapped a composi-tionally zoned magma chamber. The least evolved mag-ma occurs in the area ofthe conduit and is characterizedby relative enrichments in Si, K, Ti, Fe, Y, Sr, Th, andREE. The more differentiated magma is relatively en-riched in Li, Be, B, Na, Al, Sc, As, Rb, Sn, Cs, and Uand was erupted earlier in the volcanic cycle. Analyses oftwo glass separates show the same compositional trendsand demonstrate that these variations are shared by themelt and are not simply a function of changing crystalcontent. For most elements, there is greater range andfluctuation in concentration within the pyroclastic unitthan across the dome. In addition, some elements fluc-tuate in abundance rather than displaying a progressivechange in concentration with order of eruption. This ismore pronounced in the tephra than the dome, and weattribute this behavior to the dynamics of the eruptionby which compositionally distinct batches of magma maybe erupted simultaneously or, during a hiatus in the erup-tion, unerupted evolved magma may reoccupy the upperreaches of the magma chamber near the conduit (Spera,1984; Spera et al., 1986).

Contamination by lithic fragments from underlying la-tite or dacite is evident in the tephra sample at I I m,which has anomalously high contents of Ti, Mn, and Fe.Whole-rock F contents are variable and do not representmagmatic values; F contents in massive glass and meltinclusions range from 2.3 to 3.5o/o. The extent to whichindividual elements were concentrated upward or down-ward in the preeruptive magma chamber is revealed byenrichment factors illustrated in Figure 7, where elemen-tal concentrations in early erupted pumice are divided byconcentrations in late erupted felsite from the center ofthe dome. Those elements that were strongly enrichedupward in the magma chamber include Li, Be, B, As, Sn,Sb, Cs, Ta, and W. Strongly depleted elements arcTi,Zn,Y,Zr,Ba, REE, Hf, Pb, and Th. Ofparticular importanceare some elements that behave in opposite direction atHoneycomb Hills compared to enrichments and deple-tions in less evolved rhyolites such as the Bishop Tuff(Hildreth, 1979) or Twin Peaks, Utah (Crecraft et al.,1981). At those localities Y, Nb, and Th are enriched inevolved lavas. At Honeycomb Hills these elements aredepleted in evolved samples because of precipitation ofaccessory mineral phases containing these elements in highabundance. Christiansen et al. (1986) report a similar de-

35

30

25

20

1 5

1270

pletion of Th and LREE during evolution of a F-richgranite from western Utah. Depletions of Nb and Y werenot observed despite the presence of trace amounts ofoxides rich in Nb and Ta.

The covariance of elements with differentiation is il-lustrated in Figure 8 where concentrations are plottedagainst abundance ofRb, which exhibits substantial andsystematic variation throughout the eruptive sequence andserves as an index ofdifferentiation. The figure illustratesthe increase in abundance of Li, Be, Na, Al, As, Sn, and


4.5 5.0

1 5

1 0

5

35

30

25

20

35

30

Z J

0 200 400 150 3oO too 300 0 .5 1 .0 4 .0 50 12 14 16 72

100

I 1 0 0

20 40

Cs with differentiation and the concomitant decline inconcentrations ofSi, Fe,Zr, and all the REE. The topaz-bearing xenolith is colinear with many elements, but thereare notable exceptions such as FeO, AlrOr, andZr.

There is no fractionation of light REE with respect toheavy REE as is commonly obsewed in silicic magmasystems. All REE except Eu are depleted with differenti-ation. Chondrite normalized REE abundances produce aflat pattern (Fig. 9), characteristic of topaz rhyolites ingeneral (Christiansen et al., 1986). These data also show

200 600 1000

3 5 7

' t 5

1 0

5

0

25 40 20 40 35 60 100

Fig. 6. Elemental composition vs. stratigraphic position. Elements are in ppm; oxides are in wt0/0. The shaded region is the

approximate dome center and last erupted material. The y-axis is to scale in meters for the pyroclastic sequence but only shows

rititlu. position in the dome (above 12.5 m). Open circles are pumice. Open squares are glass separates, and filled circles are felsite.

0.75 1.25 40 80 20 40 15oo 2ooo loo 200 50 80 80 100 10 30 0

35 60 ' r0 20

Sc

Zn

" l l " o l l " " - * r *

Obserued Carcutated l l l l

\t / ll ll Enrichmentracrors

il ll ll Los (earry/rate)

Fig.7 . Enrichment factors for the Honeycomb Hills rhyolite.Early material is sample 43, and late is sample 66 [except for Be: sample 56 (late), F : sample 32G (early), and Ta (1 early; 63late)1. The comparison is for early erupted pumice vs. the centralportion of the dome, i.e., the last erupted magma. Due to vari-ability in Eu in pumice, an Eu factor has not been calculated.Calculated enrichment factors are based on 750lo crystallizationofphases in volume fraction abundances: quartz 0.330, plagio-clase 0.305, sanidine 0.333, biotite 0.030, zircon 0.0003, thorite0.000 045, fluocerite 0.0004, columbite 0.000 00 1, fergusonite0.0005, and ishikawaite 0.00005.

greater variation in REE concentration within the 12.5 mpyroclastic sequence than within the entire volume of thedome.

PanrrttoN coEFFIcIENTs

Separates of the predominant mineral phases were an-alyzed for trace element concentrations (Table 2). Figure3 illustrates the chondrite-normalized REE concentra-tions of these phases plus accessory phases analyzed byelectron microprobe. Crystal-glass partition coefrcients(K) have been calculated for each phase (Table 8). Theinclusion of accessory minerals in phenocryst phases canproduce significant and selective errors in calculated Kovalues. Of particular concern is the common inclusion ofzircon and monazite in biotite. For example, the separateofbiotite sample 32 is apparently contaminated by zirconanC monazite as revealed by high calculated partition co-efficients for Hf, U, Th, and all REE. REE-bearing acces-sory phases were not observed in biotite from pumicesample l, and we are confident that partition coefficientsfor that sample are appropriate. Zircon, thorite, mona-zite, and niobates are often included in topaz pheno-crysts, and for this reason Ku values for topaz were notcalculated. Megacrysts of topaz contain measurably lowerconcentrations of trace elements than phenocrysts, andthey are devoid of inclusions; however, the compositionof the melt from which they crystallized cannot be estab-lished. The source of the megacrysts is unknown. Theymay have grown in the Honeycomb Hills magma at anearly stage and did not incorporate accessory phases thatwere yet to nucleate, or the megacrysts may be xenocrys-tal, having crystallized earlier in a pegmatite associated

127 |

with the parent magma body of the Honeycomb Hillsmagma volume. The similarity in composition of mega-crysts and phenocrysts strongly suggests that the two aregenetically related and must have crystallized from thesame or similar magmas. Accordingly, we attribute hightrace element contents of topaz and other phenocrysts toinclusion ofaccessory phases, and we report here parti-tion coefrcients only for those phenocrysts that do nothave suspiciously high trace element contents.

Biotite selectively incorporates Sc, Zn, Co, Rb, Ta, andBa and has partition coemcients less than I for As, Cs,Hf, W, Au, Th, U, and all REE. None of the elementsanalyzed yielded partition coefficients greater than I forplagioclase, sanidine, or quartz, including Eu in feldspar.Accessory phases have extraordinarily high K" values forselected elements. Zircon and thorite incorporate acti-nides and heavy REE. Niobates may incorporate Mn, Ta,and Sn and are strongly enriched in middle REE andHREE and individually enriched in Y, IJ, and Ce. LightREE are selectively incorporated in monazite and fluoce-rite. Monazite also contains substantial Th (12.50/o). This partitioned more favorably than U in zircon, thorite,and monazite, whereas the opposite is true for niobates.

Prrvsrc,ql, PARAMETERS

Temperature

Equilibration temperatures were calculated for 64 pairsof plagioclase and alkali feldspar using the two-feldsparthermometer of Fuhrman and Lindsley (1988). These re-sults are plotted against stratigraphic position in Figure10. Temperatures are also tabulated in Table 3 for eachfeldspar pair reported there. To arrive at optimum con-vergence, the Fuhrman and Lindsley (1988) algorithmchanged the measured feldspar compositions from 0 to 2molo/o in Ab and Or in both alkali feldspar and plagio-clase. In all but three instances, An contents remained asanalyzed. The adjustments are within the precision of themicroprobe except for the positively adjusted Or contentof 0.5 to 2.0 molo/o (absolute) in plagioclase. The dataindicate overall low temperatures from approximately 570"C to 610 oC, with the suggestion amidst the scatter in thedata of a small increase in temperature with order oferuption. These temperatures are slightly lower than the600-640 "C range we reported previously (Congdon andNash, 1988) as a result of reanalysis of the entire feldsparsuite. It is not surprising that there is no large temperaturegradient. First, the erupted volume is small, and second,it is crystal rich with phenocryst contents ranging froml0 to 480/0. As Marsh (1981) has observed, the majorityof crystallization takes place within a small temperatureinterval (primarily as the result of loss of latent heat ofcrystallization).

These temperatures are lower than those of commonrhyolites but are within the range of experimental liquidiand solidi in volatile-rich silicic systems of extreme com-position such as the rhyolite of Spor Mountain, Utah, andthe glass from Macusani, Peru. At 200 MPa, the solidusfor the Spor Mountain rhyolite extends from 650 to 500


1 0

1 . 0

12',72 CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA

F

feos@

-u o@

^ q f @ o o

f o " eo

Na2O

At203

o@ o

t * o €

30

25

20

1 5

As oo f t "

@ €

@3@

Cs o

"Ee@

@ @H

1 . 1

0.9

0 .8

%c" t*o

Ce

E

%pB* oYb

30

30

20

1 5

1 0

70

60

50

40

30

E o @

@@ t r

Nd

4

30

20

1 0

0

6 @E E € ) ^ @

@ @ o t on

Lao

4

2

1

0

@B I n @

E @o -

Tb o

n oudo

e^ oSg

H

0

180160140

80

604020o

600

30

20

1 0

0

D

4

3

2

35

zc

nl 4

E

%ts'ts oLuo

l L120O 1zO0 1600 1800 2000 600 800 looo 12oo 14oo 16m 1800 2oOO 600 800 1000 1200 1400 1600 1800 2000

Rb RbRb

Fig. 8. Plots of selected elements vs. Rb. Elements are in ppm; oxides are in wt0/0. Open circles are pumice; closed circles are

domal felsite. The open square is vitrophyre 32, arrd the solid square is xenolith 10.

800

^ @ oL i . , " @ o

@

@

E @ t100

50

o

* r E @ @

E

&e o

sio2 0

o

e O@

6E r E @ a

O @

FeO* oo

E

o tooo@

@ @eS@ sn

CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA 1273

100

Rock/Chondrite

La Ce Nd Sm Eu Tb Yb Lu

Fig. 9. Chondrite-normalized REE concentrations in pumice(horizontal pattern) and dome felsite (vertical pattern).

"C depending on HrO content (Webster et al., 1987), andthe Macusanite HrO-saturated solidus is in the region of450 "C (London et al., 1988). Because we observe lavaswith up to 500/o crystals on the surface at the HoneycombHills, the magma had obviously not achieved its criticalcrystallinity where viscous flow to the surface was pro-hibited. This critical crystallinity occurs about midwaybetween the liquidus and solidus (Marsh, l98l), and it isevident that the solidus for the Honeycomb Hills magmalies below 600 "C.

Oxygen fugacity

In the absence of coexisting titanomagnetite and il-menite, ,fo, is calculated from measured FerOr/FeO inglass sample 32 by the method of Sack et al. (1980) asmodified by Kilinc et al. (1983). Using an average tem-perature of600 oC, an fo, of l0 '74 bars is calculated forthe vitrophyre, which places the magma 3 orders of mag-nitude above the fayalite-magnetite-quartz O buffer andon a trend consistent with /or-temperature conditions forsilicic ash-flow sheets from Nevada determined by Lip-man (1971).

Water fugacity

The /"ro may be determined from coexisting biotite,sanidine, and magnetite as established experimentally byWones and Eugster (1965). For the calculations presentedhere, we have adopted the equilibrium constant advo-cated by Hildreth (1977) as follows:

19973_ 1_ 0.00903 /p _ r\l l l l s z o - T

( r - 1 ,

#

{+

{#

#

zoF-ooc

uJ=k 1 0t!ct

1 0

1 . 0

0 r -540

0 .1

TEMPERATURE ( 'C)

Fig. 10. Two-feldspar temperatures vs. stratigraphic posi-tion. Error bars represent I sd about the mean. The tephra sec-tion is shown to scale in meters.

+ ln 4ft..roy5;ro16roHr2 + !"rn fo,a

- ln a$jo, - ln 4tl,r,r6r. (l)

The activity for annite is given by the ideal mixing model:

4he3Arsi3o16(ou)2:(tt2tX)(I6tXF)t(XoJ'' Q)

The activity of sanidine, Or.r, at 600 "C and 1000 barsis 0.75 (Carmichael et al., 1974). Reconnaissance micro-probe scans of magnetite show negligible amounts of Ti,and thus the activity of magnetite is assumed to be unity.At600and anfo2of l0- r?abars, fu2ordngE f rom 1000to I 135 bars in pumice and vitrophyre and from 133 to200 bars in domal lavas (Table 4). These correspond toHrO pressures of approximately 2000 and 220 bars, re-spectively (Burnham et al., 1969). The drop in /rro be-tween the pyroclastic sequence and the lavas is recordedin the increased F content and concomitant decrease inOH content of biotite.

Fluorine fugacity

Biotite phenocrysts in the pyroclastic sequence appearto have erupted rapidly, as they are homogeneous andshow no evidence ofreaction. Accordingly, they preservethe preeruptive HrO/HF ratio of the magma. In contrast,biotite phenocrysts in subsequent erupted domal lavasare enriched in F and record the passive dewatering ofthe volume of magma that was erupted to form the dome.

1274 CONGDON AND NASH: ERUPTIVE PEGMATITE MAGMA

Trele 8. Crystal-liquid partition coefficients (Ko) for phenocryst and accessory phases

Plagio- Fluo- Colum- Fergu- lshika- Cerium

cla-se Sanidine Quartz Zircon Thorite Monazite cerite bite sonite waite niobate

ScMnCoZnAsRb

ZrNbSnSbCsBaHf

WAuThULA

NdSmEUTbYbLu

19.6

2.9l e

0.13z .c

3.2u.301 . 20.882.1o.70.60.58o.770.490.590.330.620.950.s00.510.58

21.6

2.71 40.092.4

0.80.532.1

2.71 . 0

0.03

0.7o.4

0.02

0.10.01

o.220.02

0.290.120.990.420.320.14

0 . 1 10.070.07

0.03

o.70.4

0.8

0.20.030.130.140.02

o.'120.080.410 . 1 20.110.040.480.040.040.05

0.02

o.2

0.10.010.130 . 1 2

0 10.080.070.04

0 0 40.260.070.040.05

8500

7200

2501400

125560s60

30

140

1 90008000

321 258

106

320780

70

80

210

5030160

2700170030001200

1 4000oe

4000

2200

720390

6101 100

1 743

530910

300021 00990

1 000

100

6400280

: :

5400300034001 600

, :1

3700 3700540100 150

1310 1270

2600 4i7800 3100

7 7 425 790

225 113250 230

650800 960600 1390

Note.'Liquid composition is the glass,32G.

The ratio fnro/fno, which can be calculated from the an-alyzed biotite composition and experimental data of Mu-noz and Ludington (1974), is listed for representativephenocrysts in Table 4. The fugacity ratio in the pyro-clastic sequence is about 750, whereas it is about 200 inthe lavas of the dome. The significant drop in the ratioreflects preferential loss of HrO from the melt to the va-por phase, resulting in a relative enrichment in F that hasa fluid-melt partition coefficient less than one (Dingwellet al., 1985; Manning and Pichavant, 1988; Webster etal.,1987; London et al., 1988). HF fugacities range from1.4 bars in pumice and vitrophyre to 0.9 bar in domallavas (Table 4).

To assess the relative enrichment in F in micas of dif-fering Fe/Mg ratios, Munoz (1984) formulated the F in-tercept value [IV(F)], which in effect compensates for Fe-Favoidance. High relative fugacities of F result in low in-tercept values. At the Honeycomb Hills, F intercept val-ues range from 0.49 in pumice to -0.16 in the domallavas. The latter are the lowest values yet recorded ineither magmatic or hydrothermal environments. Munoz(1984) has summarized F intercept values from a varietyofsettings, including Sierran granites (2. l), rhyolites (2.0-2.3), Pikes Peak granite (1.5), Santa Rita porphyry copperdeposit (2.2-1.9), and Henderson porphyry molybdenumdeposit (0.76-0.68). The only value within the range ofthe Honeycomb Hills is 0.34 for biotite associated withSn mineralization in Nigeria. It is evident from thesecomparisons that the trIoneycomb Hills represents a mag-matic system that is extraordinarily enriched in F relativeto HrO.

Viscosity

Viscosities of dry silicic magmas are of the order ofl0'2 poise at the liquidus. If dry, the Honeycomb Hillsmagma probably would not have erupted to the surface.HrO is effective in reducing the viscosity of magmas, andthe addition of 5 wto/o HrO to a dry granitic melt servesto reduce the viscosity by 6 orders of magnitude. F servesa similar role in reducing the viscosity as well as the melt-ing temperature of silicic magmas (Mysen and Virgo,1985; Dingwell et al., 1985; Rabinovich, 1983; Wyllieand Tuttle, 1961). Dingwell et al. (1985) determined thatthe reduction in viscosity due to added F is greatest forhigh-silica rhyolites. A rhyolite melt with 750lo SiO, isreduced in viscosity from 108 poise to about 105 poisewith the addition of approximately 3 wto/o F at 1000 .C.

The effects of HrO and F appear to be additive, and Imol of F has at least the equivalent effect as I mol ofHrO (Dingwell et al., 1985). Recently, Dingwell (1987)has shown that there is a negative deviation from theadditivity of HrO and F, which further reduces viscosityin albite melts. Extrapolation of the high-temperature ex-perimental data at 7 kbar to the temperature of the Hon-eycomb Hills magma suggests that viscosities could bereduced by an additional order of magnitude over theadditive properties of the l:l molar experimental con-centrations of 5.580/o HrO and lI.7o/oF. At lower pressureand with substantially less F than in the experiment, thissynergistic effect will be less in the Honeycomb Hillsmagma but may lower viscosities slightly from those cal-culated below.

G

degassing

' \ . - 7o Crystals

crystal-free 50%

t - - \E goz

10%A D

o/oH2O + F

Fig. 11. Viscosity (poise) vs. total H,O + F content of themelt. See text for discussion.

Analyses of melt inclusions in quartz phenocrysts bysecondary ion mass spectrometry and electron micro-probe indicate HrO contents of 4.5-5.5 wto/o in the upperpart of the preeruptive magma body (Gavigan et al., 1989).Melt inclusions in domal lavas have HrO contents rang-ing from I to 3 wto/o. Similarly, F contents in melt inclu-sions decline from 3.50/o in early erupted material to 2ol0in domal lavas. Accordingly, the total HrO plus F contentin the upper portion of the magma chamber was about8-9 wto/0. In the mass of the magma that ultimately gaverise to the domal lavas, the total HrO plus F content wasapproximately 4 wto/o. Assuming that F has an effectequivalent to HrO in reducing viscosity, these volatilecontents provide viscosities for a crystal-free melt of l06 land l0e3 poise, respectively (Shaw, 1965). The rheologi-cal properties of the magma will be further modified bythe effects of crystallization and degassing, which are con-sidered below.

Crystallinity

In general, a crystal-free rhyolite melt above its liqui-dus exhibits pseudoplastic behavior (Spera et al., 1982).That is, at low strain rates the behavior is Newtonian,but at high strain rates viscosity has a negative deviationfrom linearity. However, magmas are rarely above theirliquidi, and in magmas with low volatile contents, crys-tallization tends to increase viscosity rapidly; when phe-nocryst contents exceed 50-600/0, noncreep flow is notpossible. A dramatic change in viscosity should occur ata critical crystallinity where crystal interference will causethe whole mass to lock up and freeze at depth (Marsh,l98l). A tradeoffdoes exist, however, because high crys-tal content will cause a magma to cool more slowly be-cause of the latent heat of crystallization, and a magmathat retains its heat better will remain fluid longer. Marsh(1981) places the critical crystallinity at abour 55-600/o in

1275

andesites and basalts; 5s strggests that it is substantiallylower in more viscous, silicic magmas, perhaps as low as20o/o in anhydrous melts and 25-35o/o in "moist" mag-mas. In the Honeycomb Hills rhyolites, crystallinities of500/o are approached; these levels require high volatilecontents to provide sufrciently low viscosities for flow tothe surface.

The effect of crystallization on the rheology of the Hon-eycomb Hills magma may be evaluated using Roscoe's(1953) formula:

p: pr(l - RX1-zs (5)

where p is the viscosity of the liquid plus crystals, p, isthe viscosity of the crystal-free liquid, and X is the frac-tion of crystals. R is a constant, with a value that is theinverse of the critical packing density of crystals, at whichthe mass cannot flow. To allow movement at constantvolume, a random packing density is assumed. For spheresthis is 0.62, and R : l/0.62: 1.61. For an increase inphenocryst abundance from l0o/o in early erupted pumiceto 500/o in domal lavas, the effect in an anhydrous meltis to increase viscosities by over an order of magnitudein the crystal-rich magma. Creep effects were unimpor-tant because deformed phenocrysts do not occur.

However, there is an opposing effect in volatile-richmagmas because of the concentration of volatile constit-uents in the residual melt as crystallization proceeds.Crystallization of 500/o of quartz and feldspar serves todouble the concentration of volatiles in the melt, whichreduces the effective viscosity by an amount significantlygreater than the increase produced by crystals alone. Thenet effect is that crystallization in volatile-rich systemsreduces the effective viscosity of the magma as long asthe volatiles are retained in the magma. This effect isillustrated in Figure I l. Line AC is the viscosity of crys-tal-free Honeycomb Hills melt at 600.C as a function oftotal volatile content (Shaw, 1972), assuming that F hasthe equivalent effect of HrO in reducing the viscosity ofsilicate melts. Paths AD, BE, and CF illustrate the reduc-tion in viscosity with crystallization of l0-500/o frommagma volumes with differing initial volatile contents asrecorded by melt inclusions. The decline in viscosity ofthe magma results from increased volatile concentrationsin residual liquids, whereas the deviation upward fromthe crystal-free viscosity curve results from increasedcrystal content of the magma. The curve DF representsan approximate preeruptive gradient in viscosity fromthe earliest erupted pumice, with 100/o crystals (D) to themagma volume producing the domal lavas with 500/ophenocrysts (F). Mag;ma with up to 300/o crystals vesicu-lated rapidly, fragmented, and was erupted explosively toform the pyroclastic sequence. The more crystal-richmagma degassed passively on route to the surface, withpreferential loss of HrO over F to the fluid phase thatresulted in increased viscosity along an approximate pathrepresented by FG. The path FG represents a minimumslope, because degassing on route to the surface promotescrystallization that in turn raises viscosity. This effect was


1 1

1 0

tU'osoo

(,oJ 7

1 0

r276

probably minimal because of the high proportion of Fretained by the melt. It seems probable that the eruptionceased when crystallinity in either the conduit or upperreaches of the magma chamber achieved a critical valuebeyond which flow to the surface was prohibited, regard-less of the volatile content of the interstitial melt. At theHoneycomb Hills, this value is about 500/o crystals, whichis the phenocryst content of felsite in the conduit.

The magma at point F contains 4o/oH'O as well as 40loF. In the Harding pegmatite, saturation at 4o/o HrO occursat 1000 bars pressure (Burnham and Jahns, 1962). In aSpor Mountain vitrophyre, which is compositionallysimilar to Honeycomb Hills lavas, HrO saturation occursexperimentally at about 3.30/o HrO at 500 bars and 4.5o/oHrO at 1000 bars (Webster et al., 1987). Accordingly, weestimate that passive degassing of the magma began atpressures slightly less than 1000 bars or at depths of2 to3 km. Despite the loss of HrO, the presence of substantialF kept viscosities sufficiently low to permit flow of themagma to the surface. Furthermore, if the negative de-viation from the additive properties of HrO and F ob-served in albite melts (Dingwell, 1987) occurs in naturalsilicic magmas, then an increase in the F/H'O ratio to-ward a l: I molar ratio will retard the rate of increase inviscosity that inevitably accompanies degassing. Ulti-mately, much of the F was lost during crystallization ofthe groundmass of the domal lavas.

Crrnnnrc,l.r, EvoLUTroN

Fractionation model

The data in Figures 5 and 6 illustrate that there aresystematic variations in chemical composition with re-spect to sequence of eruption and, by inference, to posi-tion in the preeruptive magma chamber. Although thelavas of the Honeycomb Hills share similar characteris-tics of enrichment factors with high-silica rhyolites, thereare distinct differences including depletion in Y, Nb, Th,and all the REE. These elements are abundant in acces-sory phases including several niobates, zircon, thorite,and fluocerite. The presence of these particular phasesand the uncharacteristic depletion of some elementsstrongly suggest that the concentrations of many traceelements in silicic magmas are controlled by crystal-liq-uid equilibria and not by diffusive properties of the meltalone. Some elements such as Sb, which are incompatiblein all phases, are simply enriched in residual melts.

The enrichment and depletion patterns observed (Fig.7) can be reproduced by Rayleigh fractionation of thephenocryst phases using partition coefficients measuredfor phenocrysts at the Honeycomb Hills. Because of thelarge number of phases present, the model has consider-able latitude and is only permissive evidence that frac-tionation is a possible mechanism for the chemical di-versity observed. The optimum result is obtained forcrystallization of 750lo of the initial magma volume: thecorrespondence of the model results with observation isillustrated in Figure 7. The abundance of phases removed


in the model is given by Congdon and Nash (1988, Fig.l). The role ofaccessory phases is evident where thoriteremoves Th, whereas U is enriched in residual liquids.Similarly, Nb and Y are removed from the magma byprecipitation of Ta-poor niobates, resulting in enrich-ment in Ta.

More than 750lo crystallization is required to explainextraordinary enrichments in B, Be, and, to a lesser extent,F, Li, and Cs. It is possible that these elements are de-pleted in the domal lavas as a result of loss to the fluidphase during degassing, so that the actual enrichmentsproduced by crystallization may not be as great as theyappear. B has a positive fluid-melt partition coefficient(Pichavant, 1981, 1987; London et al., 1988), and Web-ster et al. (1987) have shown exporimentally that Li, Be,B, and Cs are partitioned preferentially into the fluid phasein F-rich silicic melts. Accordingly, these relative enrich-ments may not reflect preeruptive gradients but are theresult of the eruption itself.

The variation in major elements illustrated in Figure 5is consistent with the differentiation of the early eruptedmagma volume in response to the crystallization of quartzand feldspar, resulting in increased F and HrO contentsin residual melts. The data are also consistent with Fcontents of up to 30lo in the early erupted magma.

As noted previously, composition varies more in thepyroclastic sequence than in the domal lavas. This maybe due in part to the ability to separate liquids from crys-tals more readily in crystal-poor systems than in crystal-rich systems where crystal movement is more inhibitedor the extraction of interstitial melt is more tortuous. Weadvocate phase separation ofa buoyant, volatile enrichedmelt at the crystallizing margin of the magma chamberand its subsequent accumulation as a density stratifiedmelt in the uppermost portion of the magma chamber. Itis evident that accessory phases are instrumental in con-trolling many trace element abundances. Fractionation ofthese phases by gravitational settling cannot be effectivebecause of their small size. Separation by inclusion inlarger phonocryst phases is possible; however, only bio-tite and topaz commonly contain accessory inclusions,and they themselves are not abundant. For this reason,we prefer a mechanism of multiphase cryslallization atthe margins of the chamber, accompanied by buoyantseparation of the residual melt fraction.

Petrogenesis

Aside from their porphyritic texture, the lavas of theHoneycomb Hills possess the chemical and mineralogicalcharacteristics of a rare element pegmatite. These bodiesare typically enriched in Be, Li, Rb, Cs, Ta, Nb, and Sn,and the concentrations of these elements at the Honey-comb Hills fall within the range reported for rare elementpegmatites (Cerni, 1982). For example, Sn is present inconcentrations up to 33 ppm at the Honeycomb Hills;granites with Sn mineralization are reported to contain16 to 32 ppm Sn, whereas 3 ppm is usual for commongranite (Wedepohl, 1970). REE-bearing accessory phases

such as fluocerite and niobates are common in rare ele-ment pegmatites (Henderson, 1984). Pegmatites com-monly contain two micas. Muscovite phenocrysts are rarein volcanic rocks, and at the Honeycomb Hills topazmayproxy for muscovite by the following reaction, which hasa positive molar volume change and would be favoredby lower pressures:

KAlrAlSi3O,0(OH,F), + SiO, : AlSiOo(OH,F),muscovite quartz topu

+ KAlSi3O8.sanidine

Domal lavas contain pegmatite xenoliths as well as mega-crysts of sanidine, topaz, and biotite up to 8 cm in di-ameter. All these characteristics indicate that the lavas ofthe Honeycomb Hills were erupted from a pegmatiticmagma.

The origin of the pegmatitic magma remains moreenigmatic. We envision it arose by prolonged differenti-ation within a cupola in the upper reaches of a graniticmagma whose own origin provided modest enrichmentsin rare elements and F. The most likely source region forsuch a magma lies in the Proterozoic continental crust.Christiansen and ke (1986) surveyed the distribution ofF in granitoids of the western United States and dem-onstrated that the highest F contents were in Precambriangranitoids. We agree with their suggestion that F-richmagmas could be produced from these protoliths by an-atexis of Precambrian terrains of Protoerozoic age inwhich the F is derived from the breakdown of biotitewhose F contents have been enriched through one or moreprior thermal events.

AcrNowr,nlcMENTS

Support for this study was provided by National Science Foundationgant EAR-8720627 to W.P.N. and the Mineral Leasing Fund of theUniversity of Utah. The Cameca electron microprobe was purchased withfunds from the National Science Foundation (grant EAR-8618101) andthe University of Utah. The manuscript was substantially improved bythe thorough reading and many helpfirl comments of Eric H. Christiansen.

RrrnnnNcBs crrnnBacon, C.R, Macdonald, R., Smith, R.L., and Baedecker. p.A.0981)

Pleistocene high-silica rhyolites of the Coso volcanic field, Inyo Coun-ty, California. Journal of Geophysical Research, 66, 10223-lo24l

Burnham, C.W., and Jahns, R.H. (1962) A method for determining thesolubility of water in silicate melts. American Journal of Science, 260,721-7 45.

Burnham, C.W., Holloway, J.R., and Davis, N.F. (1969) Thermodynamicproperties ofwater to 1000 eC and 10,000 bars. Geological Society ofAmerica Special Paper 132,96 p.

Carmichael, I.S.E., Turner, F.J., and Verhoogen, J. (1974) Igneous pe-trology, 739 p. McGraw-Hill, New York.

Cern!, P. (1982) Petrogenesis of granitic pegmatites. In P. Cern!, Ed.,Granitic pegmatites in science and industry. Mineralogical Society ofCanada Short Course Handbook. 8. 405-462.

Christiansen, E.H., and ke, D.E. (1986) Fluorine and chlorine in grani-toids from the Basin and Range province, western United States. Eco-nomic Geology, 81, 1484-1494.

Christiansen, 8.H., Bikun, J.V., Sheridan, M.F., and Burt, D.M. (1984)

r277

Geochemical evolution oftopaz rhyolites from the Thomas Range andSpor Mountain, Utah. American Mineralogist, 69,223-236.

Christiansen, E.H., Sheridan, M.F., and Burt, D M. (1986) The geologyand geochemistry ofCenozoic topaz rhyolites from the western UnitedStates. Geological Society of America Special Paper 2O5, 82 p.

Congdon, R.D., and Nash, W.P. (1988) High-fluorine rhyolite: An erup-tive pegmatite magma at the Honeycomb Hills, Utah. Geology, 16,1018 -1021 .

Crecraft, H.R., Nash, W.P., and Evans, S.H., Jr. (1981) Late Cenozoicvolcanism at Twin Peaks, Utah: Geology and petrology Journal ofGeophysical Research, 86, 10303-10320.

Deer, W.A., Howie, R.A., and Zussman, J. (1962) Rock-forming miner-als, vol. 3, 270 p. Wiley, New York.

Dingwell, D.B. (1987) Melt viscosities in the system NaAlSi.O*-HrO-F,O ,. Geochemical Society Special Publication No. l, 423-432.

Dingwell, D.B., Scarfe, C.M., and Cronin, D.J. (1985) The effect of fluo-rine on viscosities in the system NarO-Al,Or-SiO,: Implications forphonolites, trachytes and rhyolites. American Mineralogist, 70, 80-87.

Drake, M.J., and Weill, D.F. (1975) Partition of Sr, Ba, Ca, Y, Eu*', Eu*3,and other REE between plagioclase feldspar and magmatic liquid: Anexperimental study. Geochimica et Cosmochimica Acta, 39,669-712.

Fuhrman, M.L., and Lindsley, D.H. (1988) Temary feldspar modelingand thermometry. American Mineralogist, 7 3, 2Ol-215.

Gavigan, T.H., Nash, W.P., and Webster, J.D. (1989) Preeruptive gra-dients in silicic magma as recorded by melt inclusions. Eos, 70, 1417.

Hagstrum, J.T., and Gans, P.B. (1989) Paleomagnetism of the OligoceneKalamazoo Tuff: Implications for middle Tertiary extension in eastcentral Nevada. Journal of Geophysical Research, 94, 1827-1842.

Henderson, P. (1984) Rare earth element geochemistry, 510 p. Elsevier,New York.

Hildreth, E.W. (1977) The magma chamber of the Bishop Tuff Gradientsin temperature, pressure, and composition, 328 p. Ph.D. thesis, Uni-versity of California, Berkeley, California.

-(1979) The Bishop Tuff Evidence for the origin of compositionalzonation in silicic magma chambers. Geological Society of ArnericaSpecial Paper 180, 43-75.

Hinue, L.F. (1988) Geologic history of Utah. Brigham Young UniversityGeology Studies Special Publication 7,202 p.

Hogg, N.C. (1972) Shoshonitic layas in west-central Utah. Brigham YoungUniversity Geology Studies, 10, 133-184.

Kilinc, A., Carmichael, I.S.E., Rivers, M.L., and Sack, R.O. (1983) Theferric-fenous ratio of natural silicate liquids equilibrated in air. Con-tributions to Mineralogy and Petrology, 83, 136-140.

Kovalenko, V.I. (1973) Distribution offluorine in a topaz-bearing quartzkeratophyre (ongonite) and rnobility of fluorine in quartz melts. Geo-chemistry International, 10, 4 l-49.

Kovalenko, V.I., and Kovalenko, N.I. (1976) Ongonites (topaz-bearingquartz keratophyre)-Subvolcanic analogue of rare metal Li-F granites.Nauka Press, Moscow (il Russian).

Kovalenko, V.I., Kuz'mun, M.I., Antipin, V.S., and Petrov, L.L. (1971)Topaz-bearing quartz keratophyre (ongonite), a new variety of subvol-canic igneous vein rock. Doklady Akademii Nauk SSSR, Earth SciencesSection.199. 132-135.

Kovalenko, V.I., Antipin, V.S., Konusova, V.V., Srnirnova, Ye.V., Pe-trov, L.L., Vladykin, N.V., Kuznetsova, A.I., Kostukova, Ye.S., andPisarskaya, V.A. (1977) Partition coefficients offluorine, niobium, tan-talum, lanthanum, ytterbium, tin and tungsten in ongonite. DokladyAkademii Nauk SSSR, Earth Sciences Section, 233, 2O3-2O5.

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

Lipman, P.W. (1971) Iron-titanium oxide phenocrysts in compositionallyzoned ash-flow sheets from southern Nevada. Joumal of Geology, 79,438-448.

London, D., Hervig, R.L., and Morgan, G.8., IY (1988) Melt-vapor sol-ubilities and elemental partitioning in peraluminous granite-pegmatitesystems: Experimental results with Macusani glass at 200 MPa. Con-tributions to Mineralogy and Petrology, 99,360-373.

Manning, D.A.C. (1981) The effect of fluorine on liquidus phase relation-ships in the system qz-ab-or with excess water at I kb. Contributionsto Mineralogy and Petrology, 76,206-215.

Manning, D.A.C., and Pichavant, M. (1988) Yolatiles and their bearing


r278

on the behayiour of metals in granitic systems. Canadian Institute ofMining and Metallurgy Special Yolume 39, 13-24.

Marsh, B.D. (1981) On the crystallinity, probability of occurrence andrheology of lava and magma. Contributions to Mineralogy and Petrol-ogy, 78, 85-98.

McAnulty, WN., and kvinson, A.A (1964) Rare alkali and berylliummineralization in volcanic tuffs, Honey Comb Hills, Juab County, Utah.Economic Geology, 59, 7 68-7 7 4

Munoz, J.I. (1984) F-OH and CI-OH exchange in micas with applicationsto hydrothermal ore deposits. In Mineralogical Society of America Re-views in Mineralogy, 13, 469-493.

Munoz, J.I., and Ludington, S.D. (1974) Fluoride-hydroxyl exchange inbiotite. American Journal of Science. 27 4. 396-413.

Mysen, 8.O., and Virgo, D. (1985) Structure and properties of fluorine-bearing aluminosilicate melts: The system SiO,-F at I atm. Contribu-tions to Mineralogy and Petrology, 91, 205-220.

Nash, W P., and Crecraft, H.R (1 985) Partition coefficients for trace el-ements in silicic magmas. Geochimica et Cosmochimica Acta , 49, 2309-2322.

Pichavant, M. (1981) An experimental study ofthe effect ofboron on awater saturated haplogranite at 1 kbar pressure. Geological applica-tions Contributions to Mineralogy and Petrology, 76,430-439.

-(1987) Efects of B and HrO on liquidus phase relations in thehaplogranite system at I kbar. Bulletin Mineralogique, 106,2Ol-211.

Pichavant, M., and Manning, D.A.C. (1984) Petrogenesis of tourmalinegxanites and topaz rhyolites; the contribution of experimental dataPhysics ofthe Earth and Planetary Interiors, 35, 31-50.

Pichavant, M., Herrera, J.V., Boulmier, S., Briqueu, L., Joron, J.-L., Ju-teau, M., Marin, L., Michard, A., Shepard, S M.F., Treuil, M., andVernet, M. (1987) The Macusani glasses, SE Peru: Evidence of chem-ical fractionation in peraluminous magmas. Geochemical Society Spe-cial Publication. l. 359-373.

Rabinovich, E.M. (1983) On the structural role offluorine in silicate glass-es. Physics and Chemistry ofGlasses, 24,54-56.

Roscoe, R. (1953) Suspensions. In J.J. Hermans, Ed., Flow properties ofdisperse systems, p. 1-38. North Holland, Amsterdam.


Sack, R.O., Carmichael, I.S.E., Rivers, M., and Ghiorso, M.S. (1980)

Ferric-ferrous equilibria in natural silicate liquids at I bar. Contribu-tions to Mineralogy and Petrology, 7 5, 369-376

Shaw, H.R, (1965) Comments on wiscosity, crystal settling, and conYec-tion in granitic magmas. American Journal ofScience,263, 120-152.

- (1g/2)Yrscosities of magmatic silicate liquids: An empirical meth-od of prediction. American Journal of Science, 27 2, 87 0 -89 3.

Spera, F.J. (1984) Some numerical experiments on the withdrawal of mag-ma from crustal reservoirs. Journal ofGeophysical Research, 89, 8222-8236.

Spera, F.J., Yuen, D A, and Kirschvink, S.J. (1982) Thermal boundarylayer convection in silicic magma chambers: Effects of temperature-dependent rheology and implications for thermogravitational chemicalfractionation. Joumal of Geophysical Research , 87 , 87 55-87 67 '

Spera, F.J., Yuen, D.A., Greer, J.C., and Sewell, G. (1986) Dynamics ofmagma withdrawal from stratified magna chambers. Geology, 14,723-I Z O -

Turley, C.H., and Nash, W.P. (1980) Volcanism in western Juab andMillard counties, Utah. Utah Geological and Mineral Survey SpecialStudies. 52. 1-33.

Webster, J.K.D., Holloway, J.R., and Hervig, R.L. (1987) Phase equilbriaof a Be, U and F enriched vitrophyre from Spor Mountain, Utah.Geochimica et Cosmochimica Acta, 5 l, 389-402.

Wedepohl, K.H. (1970) Tin. Handbook of geochemistry. Springer-Verlag,New York.

Wones, D.R., and Eugster, H.P. (1965) Stability of biotite: Experiment,theory and application. American Mineralogist, 50, 1228-127 2.

Wyllie, P.J., and Tuttle, O.F. (1961) Experimenlal investigation of silicatesystems containing two volatile components-part IL The effects ofNH, and HF, in addition to HrO on the melting temperatures of albiteand granite. American Journal of Science,259, 128-143.

Mnruscnrrr REcErwD Mancn 12, 1990MaNuscnrrr AccEmED Apnn 15, 1991

eruptive pegmatite magma: rhyolite of the honeycomb hillso ...american mineralogist, volume 76,...

Documents