American Mineralogist, Volume 77, pages 314-328, 1992

Chemical controls on ferrierite crystallization during diagenesis of silicicpyroclastic rocks near Lovelock, Nevada

SrBpnBN B. Rrcn*Exxon Research and Engineering company, Annandale, New Jersey 08801, U.S.A.

Knrrn G. PIPrB2180 Schroeder Way, Sparks, Nevada 89431' U.S.A.

Devro E. W. Ylucn.lNExxon Research and Engineering company, Annandale, New Jersey 08801, u.s.A.

Ansrru'cr

The zeolites ferrierite, mordenite, and clinoptilolite have formed with smectite through

hydration reactions in a sequence of rhyolitic pyroclastic rocks near Lovelock, Nevada.

Fiigh activities of silica and magnesia were required simultaneously for the large-scale

crystallization of ferrierite in these diagenetically altered rocks. Post-depositional addition

of Mg coincided with ferrierite and smectite formation. The source of the Mg is not known,

but it could have originated either from a brackish, saline lake or from nearby leached

basalts. Thus, origins of the Lovelock ferrierite deposit may have involved both lacustrine

deposition and an open ground water system for the transformation of vitric tuffs into

zeolites.Ferrierite crystallized early, mainly in the pore space of the tuffs. Clinoptilolite and

cristobalite grew later by replacement of glass shards. Orthoclase and mordenite crystal-

lized somewhat later, orthoclase usually in intimate association with ferrierite in K-rich

zones and mordenite alone in open pore spaces. Ferrierite crystallization proceeded through

a mechanism of dissolution and reprecipitation on glass and clay surfaces' Significant

variation in chemical composition of individual ferrierite crystals occurs on the scale of a

few cubic millimeters.

INrnonucrroN

Ferrierite is a medium-pore zeolite that has interestingpotential applications in petrochemical processing.Klowledge of the conditions leading to its formation mayprovide insight into creating a simple, inexpensive syn-thesis method. The Lovelock, Nevada, deposit, in whichferrierite occurs in minable proportions, provides a nat-ural laboratory for studying the processes of zeolitizationand the conditions under which ferrierite crystallizes.What were the special conditions responsible for the un-usually large quantity offerrierite at Lovelock? From theresults of a study of ferrierite synthesis, Cormier and Sand(1976) concluded that metastability may be the key. Theirconclusion, although soundly based on experiment andintuitively pleasing when applied to the Lovelock depos-it, nevertheless requires further examination throughcareful study of the mineralogical and chemical relation-ships in the rocks themselves. In this study we have com-bined electron petrographic characterization with bulkchemical and phase assemblage data to interpret the or-igins offerrierite in altered pyroclastic rocks.

* Present address: Exxon Production Research Company, P.O.Box 2189. Houston. Texas 77252-2189, U.S.A.

Since its discovery in an outcrop ofolivine basalt near

the shore of Kamloops Lake in British Columbia (Gra-

ham, l9l8), the zeolite ferrierite has been observed in

nearly 30 localities. With the exception of a very fewoccurrences in altered volcanogenic sediments, these allappear to be hydrothermal vug occurrences. In the de-posit near Lovelock, discovered in 1965 by P' E' Galli,ferrierite formed through hydration reactions in a se-quence of rhyolitic pyroclastic rocks. A plzzle at Love-lock is the dominance of ferrierite over mordenite and

clinoptilolite, two zeolites that are very common ln slm-ilar tuffaceous settings in the western United States. Infact, several zeolites, including erionite, mordenite, chab-azite, clinoptilolite, and analcime, are very common in

such deposits, whereas ferrierite is only known to occurin the Lovelock deposit and in smaller amounts in theStillwater Range near Fallon, Nevada. A Korean occur-rence was interpreted by Noh and Kim (1986) to resultfrom diagenesis of silicic vitric tuffs.

Reports describing the Lovelock zeolite deposit con-tain widely divergent opinions about the genesis of thezeolites. Sand and Regis (1968) suggested a hydrothermalorigin for the ferrierite, but no supporting evidence forestimates of temperature or sourc€ of the fluids was given.

Regis (1970) described the Lovelock and Stillwater Range

0003-004x/92l0304-03 I 4$02.00 314

RICE ET AL.: FERRIERITE CRYSTALLIZATION 3 1 5

a Ferrierite Deposit

Fig. 1. L,ocation map for the Lovelock zeolite deposit (Shep-pard et al., 1983).

occurrences of central Nevada, giving compositions ofthe ferrierite at each locality, and he also reported zona-tion for ferrierite, mordenite, and clinoptilolite based onoutcrop sampling. Low alkalinity conditions were sug-gested. In a guidebook accompanying the zeolite fieldtripZeo-Trip'83, Sheppard et al. (1983) described the Love-lock occurrence in somewhat $eater detail and concludedthat the original pyroclastic material was deposited in alake but that little was known about the fluids responsiblefor zeolitization. An examination of the chemical con-trols on zeolite crystallization in the Lovelock depositthat account for this unusual occurrence is the subject ofthe present study.

B.q.crcnouNn

Geologic setting of the Lovelock deposit

The Lovelock zeolite deposit occurs in a group ofun-named sedimentary and volcanic rocks of Tertiary age(Miocene or Pliocene) in a small valley within the TrinityRange (T. 28 N., R. 30 E.) in Pershing County, northwestNevada (Fig. l). Reconnaissance mapping placed theLovelock deposit in the Upper Tertiary volcanic group(Tuv) of Stewart and Carlson (1978). The ferrierite de-posit of the Stillwater Range also lies within the Uppergroup. The observable surface features of the deposit andthe core locations are shown in Figure 2.

The stratigraphy based on the bulk chemistry, petrog-raphy, and phase-assemblage data is shown in Figure 3.There is a zonation characteizedby mordenite at the top,grading downward into less mordenite and increasingamounts of ferrierite, and finally a large amount of glassy

I Ferrieriterich tulf

Z Mordenit+rich tuff

tg zeolitic tuff

N ctass

I Diatomite

n welded rhyolite tufl

E Alluvium

Fig.2. Geologic map of the Lovelock deposit.

material. This sequence agrees with the order of appear-ance of these zeolites deduced from the experiments ofCormier and Sand (1976). The cores themselves, how-ever, are more difficult to correlate. The discontinuitiesbetween cores 2. 3. and 4 could be fault-bounded discon-tinuities or unconformities related to variations in thesedimentary environment such as erosional surfaces orpinch-outs.

Cores and outcrop samples

Four vertical core holes (Fig. 2), each 5 cm (2 in.) indiameter, were drilled under contract for W. R. Graceand Company in the early 1970s. Portions of the cores

TJ U m

Vedical EraggeEtion-10x

Mordenile Hill

A welded ftyolite tutl

D Fine gla$ + smectite

E coaR gla$ + Eolile

N Ferrlerhe + smecllte

I Benlonile

Z Mordenite + lerrierite + orthoclag

I Ferrierite + orthoclase

I Oinoptilollre + crlstoballte

Fig. 3. Stratigraphy ofthe Lovelock zeolite deposit. LayersA, B, and C are defined on the basis of composition (as in Fig.I 4) and assemblage.

foE-l I ptiocene toSedimentary and rclcanic rocks I Hotocene

I T$ I mZ i ptircene andTutf and sedimentary rocks Basalt and andesite I Miocene

GTn;l 1 Miocene toVotcanic rocks J Oligocene

tl:KE,ll I CretaceoustoMetamorphic and plutonic rocks J Triassic

OUATERNARY

TERTIARY

CRETACEOUS toTRIASSIC

LOVELOCK ZEOLITE DEPOSIT

-ll6mts

a Core Hole Locations

\l l

a

3 1 6 RICE ET AL.: FERRIERITE CRYSTALLIZATION

Fig. 4. SEM micrograph showing (a) a fracture section ofglass in sample SR75 and (b) the morphology of smectite on theglass surface.

are stored at the Nevada Bureau of Mines and Geologyin Reno, Nevada. Samples were selected from the fourcores, with particular attention paid to the best preservedcores from hole nos. 2 ar;,d 3. Unfortunately, the lowercore from hole no. 3, below about 12.2 m, was inadver-tently discarded at W. R. Grace and Company.

Vitric and zeolitic tuffs were sampled in outcrop toaugment the core samples. The ferrierite-rich outcropswere surprisingly difficult to sample because of their ex-treme toughness. They have a green color that is subtlydifferent from the other outcrops of tuffs, most of whichare buff to tan. They are also locally strongly silicified,although the spatial relationship between the opaline zonesand the ferrierite tuff was not clear in the available out-crops. There is banding in many of the zeolitic tuffs, andsome sedimentary features, such as ripple marks, pinch-outs, and cross-bedding relationships, were preserved de-spite the zeolitization process.

Instrumental techniques

Powder X-ray diffraction (PXD) was carried out usingpowder-pack methods with an automated Siemens D500

diffractometer with CuKa radiation. Scans were run al2per min from 2 to 60 20, at 40 kV and 30 mA. Phaseidentification, particularly for zeolites, was enhanced bya computer-based matching progam with a library ofnatural and synthetic zeolites as well as other silicates.

The scanning electron microscope used was a JEOL35C equipped with an LaB6 gun, a four-crystal wave-length spectrometer (WDX), adjunct energy-dispersiveX-ray detector (EDX), and PGT System IV analysis sys-tem. Both normal secondary electron imaging (SEI) andbackscattered electron imaging (BEI) were used. Electronmicroprobe analysis was performed by a method com-bining EDX and WDX, using orthoclase from BensonMines, St. Lawrence County, New York (K, Al, Si), lab-radorite plagioclase from l,ake County, Oregon (Ca, Al,Si), and hornblende from Kakanui, New Zealand (Mg,Ti, Fe, Al), as standards. Alkali difusion during analysisof zeolites and glass was corrected by monitoring countsfor 30 s and extrapolating to zero time. Clay and zeoliteanalyses totaled less than l00o/o primarily because of thepresence of HrO, both adsorbed and structural. Ferrieritetypically has l0-l5o/o HrO by weight. Matrix correctionswere calculated by the NBS Frame C program. The sec-ondary ion mass spectrometer (SIMS) used was a Cameca3f with a high-gain camera (20 million ASA) and a Crys-tal image processor (kta, 1985). Because the chargingproblems inherent with the imaging SIMS method aremore severe than with SEM, SIMS samples required em-bedding in a conductive mount.

A Philips 420ST transmission electron microscope(TEM) with EDX and adjunct PGT System IV analysissystem was used for characterization ofglassy and zeolitictuffs and for identification ofzeolites by electron diffrac-tion. Chemical compositions of ferrierite crystals weredetermined using the Cliff-Lorimer thin-film method(Goldstein et al., 1977).

Rnsur-rsElectron petrography

Vitric tufls. Vitric tufffrom 30.5-m depth in hole no.3 is relatively unaltered. Glass constitutes most of thespecimen, and the remainder is quartz, plagioclase, andphenocryst sanidine and Fe-Ti oxides. These are clearlyair-fall tuffs because, in addition to good sorting (averageshard size is 50-100 prm), there is a large amount of in-terclast porosity and no evidence for either welding orflow.

A fine-grained unaltered ash (SR73) outcrops at thebase of the welded tuff at the northern edge of the fieldarea. These are angular shards exhibiting elongated poresand occasional cuspate morphology and whose averagegrain size (5-10 pm) is very similar to the Hekla Ho teph-ra layer. Heiken and Wohletz (1985) interpreted those tobe of phreatoplinian origin, implying a highly vesicularmagma resulting from mixing with surface HrO.

Figure 4 shows fresh, fractured glass shards (50-100prm grain size) from SR75 and SR76 with smectite rims,which have this feature in common with C3-30.5. Smec-

RICE ET AL.: FERRIERITE CRYSTALLIZATION 5 t I

tite, the predominant authigenic phase in the early stagesof alteration, is present on most of the glass surfaces ob-served. The morphology shown in Figure 4b is typicallyobserved for smectite, althqugh at least in some in-stances, this corn-flake texture may be an artifact ofthepreparation for SEM.

SIMS images from polished sections of specimen C3-30.5 in Figure 5 illustrate several salient features oftheseslightly altered shards. The K ion image indicates thatalkalies have not been leached from this glass, a conclu-sion also inferred from the composition of the glass (Ta-ble l) because Na (3.020lo NarO) and K concentrations(5.25o/o KrO) are typical ofrhyolites with about 780/o SiOr.The Mg image and corroborating TEM examinationclearly show that Mg is concentrated in the smectite phaserimming the glass. Therefore, the glass alteration hereappears to be limited to the surface of the shards.

Cristobalite, clinoptilolite, and possibly amorphous sil-ica are pseudomorphic after shards (Fig. 6) and thus formin replacement reactions in several hole 2 tuffs. Ferrieriteis present in only minor amounts and occurs as coatingson shards. Figure 7a shows in cross section a relict pum-ice shard that is now largely cristobalite. Much of thesilica reported in this study is probably the opal-CT phasedefined by Jones and Segnit ( I 97 I ), characterized by lepi-spheres composed of intersecting platelets of disordereda cristobalite. Because the 4.3-A tridymite peak charac-teristic of opal-CT was not always observed, we refer toit throughout as cristobalite. The pores, which appearnearly circular in this view, are a composite of authigenicorthoclase and ferrierite, lined with smectite. The ortho-clase crystals, distinguishable in powder X-ray patternsfrom sanidine phenocrysts, form an elliptical pattern thatsuggests a concentration gradient of K with ellipsoidalsymmetry. Figure 7b shows a void outlined by smectiteand ferrierite in succession. It is not known whether thereis any direct relationship between the zeolite and the clayin terms of nucleation, crystallographic control, chemicalexchange, or reaction. Probably the clay merely provideda substrate on which the ferrierite could grow becauseferrierite was also observed to have grown directly onremnant glass surfaces in specimens C2-3.7 and C4-8.5.

Products occur at shard edges and in pores in the hole-2and hole-4 tuffs, which have carried the alteration processfurther than the early smectite alteration stage represent-ed by specimen C3-30.5. These additional crystals at-tached to shards (Fig. 8a) have a morphology and com-position confirming their identity as ferrierite. They appearphysically separated from glass by a clay layer. Ferrieritetlpically develops in a prismatic habit normal to the shardsurfaces (Fig. 8b). There are very large, highly vesicularpumiceous grains and blocky, nonvesicular shards. Per-litic texture is very common in some of the blocky non-porous particles.

Some bentonitic beds are interspersed with glassy andzeolitic tuffs. In certain strata the smectite is pure enoughto be considered for recovery, along with the zeolite, werethe deposit to be mined commercially.

Fig. 5. SIMS images of a polished section of vitric tuffsam-ple C3-30.5. The Si image identifies the shape ofthe glass shard.The Na image, which imitates Si here, confirms that alkalies areuniformly present in the glass and therefore leaching has notoccurred. The Mg is a constituent of the smectite coating theglass.

3 1 8

Tlale 1. Glass compositions

RICE ET AL.: FERRIERITE CRYSTALLIZATION

sio,Al2o3MgoCaONarOK"OTio,FeO

Total

Remarks

Sample no.

75.01 1 . 80 0 0o.222.885.130.000.00

95.02

c3-30.5

74.21 1 . 90.000.172.834.860.000.00

94.00

c3-30.5

73.51 1 . 80.000.282.715.040.000.00

93.40

c3-30.5

74.412.20.00o.223.275.100.000.00

95.12

c3-30.5

77.112.90.000.004.605.760.000.00

100.35

vesicular

sR78

74.8't2.20.000.314.594.400.000.00

96.25

vesicular

SR78

76.212.40.050.005.334.490.000.00

98.48

75.712.40 .140.894.124.260.000.00

97.51

nonvesic. nonvesic.

SR78 SR78

Ferrierite tuffs. These tuffs have been converted to as-semblages of ferrierite, authigenic orthoclase, cristobalite,and minor mordenite. Some of these tufs are nearly pureferrierite. The chemical compositions of several ferrieritesamples from Lovelock are given in Table 2. Like otherferrierite samples (Wise and Tschernich, 1976), these aremagnesian. The tuffaceous ferrierite samples from theLovelock, Stillwater, and South Korean occurrences,however, have distinctively high K concentrations.

There are several manifestations of the preservation ofthe microscopic vitroclastic texture. One of these takesthe form of ferrierite and authigenic potassium feldsparin a dense, fine-grained intergrowth (Fig. 9) in which theorthoclase inhabits the periphery of former shards. An-other feature is the presence throughout the tuffs ofshard-shaped voids. It is not always obvious whether particularvoids represent primary porosity or secondary porosityintroduced by the dissolution of glass. Those voids withdefinite cuspate pumice morphologies appear to be dis-solved shards. That early ferrierite crystals grew into pri-mary pores is supported by the observation in core 4 tuffsof ferrierite crystal egowth on fairly fresh glass shards.Where ferrierite was permitted to crystallize in an uncon-strained space the morphologies represented in Figure l0were observed. A third feature is the occurrence ofcris-tobalite, which is commonly found as spherical polycrys-tals, often seen along or adjacent to remnant vitric out-lines. In Figure 6 the cristobalite spheres are just visiblewithin the limbs of the relict shards. The K SIMS imagein Figure I l, which clearly distinguishes the orthoclase

TABLE 2. Ferrierite chemical analyses

from the ferrierite, also shows the vitroclastic texture.Phenocryst sanidine contains significant concentrationsof Na, whereas the authigenic orthoclase contains almostnone (Table 3). This allows the two feldspars to be easilydifferentiated by SIMS or EDX analysis when the grainsize or textures are not definitive.

Mordenite tuffs. Although most zeolitic tuffs containboth ferrierite and mordenite, the predominance of mor-denite in some tuffs made textural distinctions possible.Close inspection of Figures I2a and l2b reveals a fine,lacelike network of authigenic orthoclase and cristobalite,apparently positioned at former rims of glass shards. Thismicrotexture was also observed in the ferrierite tuffs, butthe contrast between zeolite-filled void space and the nar-row network of orthoclase and cristobalite was not so welldeveloped as in the mordenite-rich tuffs. Note also thepresence in Figure l2b of a double-walled network main-ly composed of orthoclase. The space between the wallsappears to be void. The mordenite tuffs are in generalmuch more porous than ferrierite-rich tuffs because themordenite grows, with few exceptions, as bundles andradial sheaves of prismatic crystals. These (e.g., Fig. l2a)are morphologically indistinguishable from acicular fer-rierite. In contrast to the intimate intergrowth of ferrieritewith orthoclase, in which the individual crystallites areoften not discernible by SEM, no similar textures wereobserved for mordenite.

Typical Lovelock mordenite and clinoptilolite com-positions are given in Table 4. The somewhat unusualcomposition for the clinoptilolite (high Si/Al and high

sio,Al2o3MgoCaONaroKrOTio,FeO

Total

Remarks

Sample no.

69.710 .60.48'1.74

1 . 1 84 . 1 10.40o.82

89.02

matrix

KP52

69.910.31 .980.990.304.120-010.25

87.90

matrix

KP52

69.310.92.511 .46o.204.010.270.73

89.45

matrix

KP52

72.911.43 .131 .060.593.640.451 .36

94.55

vug

KP52

70.410 .01 .610.610.015.050.040.23

87.89

matrix

SR64

70.1' t2.1

1.950.740.015.460.000.45

90.80

matrix

SR64

76.21 1 . 61.742.780.001 .700.701 .41

96.1 4

c2-14.3

73.51 1 . 31 .892.890.081 .970.681 .45

93.71

c2-'t4.3

73.7 69.61 1 . 3 1 1 . 12j9 2.482.10 ' l .72

0.30 0.302.56 2.960.48 0.2'l't.21 0.66

93.83 89.01

c2-17.4 C2-17.4

RICE ET AL.: FERRIERITE CRYSTALLIZATION 3r9

Fig. 6. Backscattered electron image of a cross section ofsample C2-3.7 showing replacement of glass by clinoptilolite (CL)and cristobalite (CR). Ferrierite (FER) and orthoclase (OR)formed between the limbs of the shard. Some of the ferrieritehas grown normal to the former glass surface. Note also thatcristobalite spheroids occur as discrete crystallites within thelimbs.

Ca) makes a note concerning heulandite group nomen-clature worthwhile. The zeolites heulandite and clinopti-lolite have the same framework structure and are distin-guished on the basis of composition [heulandite if Ca >(Na + K) according to Mason and Sand, 1960, or if Si/Al < 4 according to Boles, 19721 or structural stabilityupon heating to 450 "C (heulandite if the structure col-lapses, according to Mumpton, 1960). Our phase (Table4, analyses 3 and 4) could be termed heulandite on thebasis of cations because Ca > (Na + K) or clinoptiloliteon the basis of framework Si/Al ratio (-4.8). We did notperform the heating test. We call this phase clinoptilolitebecause heulandite rarely has a sedimentary occurrence,whereas clinoptilolite and mordenite have a frequent as-sociation in altered tuffs (e.g., Gottardi and Galli, 1985),particularly in high-silica environments.

Paragenetic sequence

One of the first changes attending glass alteration atLovelock was the crystallization of smectite rich in Mg,Ca, and Fe as an abundant shard-rimming phase (e.g., inspecimens C3-100 and SR74-76). The appearance of clay,and in particular smectite, is a frequently observed phe-nomenon in altered glasses of a range of compositions.Smectite was reported, for instance, by Broxton et al.(1987) in altered rhyolitic glasses at Yucca Mountain, Ne-vada; by Keller et al. ( 197 I ) in altered rhyolite near a hotspring in Michoacan, Mexico; and by Boles and Coombs(1975) as the first phase ofalteration in Triassic tuffs ofandesitic through rhyolitic composition. Davies et al.(1979) reported smectite-coated glass grains in the firstalteration zone of Guatemalan andesitic volcaniclastics.Sheppard and Gude (1968) reported smectite as a rem-nant phase adjacent to altered glassy shards in the tuffs

Fig. 7. (a) Back-scattered electron image of sample C2-17 '4

showing orthoclase (OR), ferrierite (FER), and smectite (SM)

filling the pores of a relict shard. The shard walls are the con-nected curvilinear features separating rounded pores. Some ofthis wall material is now cristobalite (CR). (b) Fracture sectionof sample C2-17.4 showing empty pore space (black) rimmed bysmectite and ferrierite.

of Lake Tecopa, a saline, alkaline lake. Other examplesinclude Boles and Surdam (1979), Brey and Schmincke(1980), and Hay (1963). Smectite may not always precede

zeolitization, but it clearly is an early stage ofhydration

TABLE 3. Feldspar chemical analyses

b

sio,Al,o3MgoCaONaroK"OTio,FeO

Total

Type

Sample no.

68.715.3o.280.810.25

13.20.000.00

98.41

orthoclase

KP52

69.118.70.800.320.00

1 1 . 10.000.00

100.00

orthoclase

c2-'t4.3

72.1 69.314 .8 18 .00.68 0.000.58 0.440.23 3.10

11.2 8.670.00 0.010.00 0.65

99.62 100.18

orthoclase sanidine

c2-17A SR3

320 RICE ET AL.: FERRIERITE CRYSTALLIZATION

Fig. 8. (a) Cross section of a single glass shard (G) with fer-rierite at its periphery. 0) Higher magnification image of fer-rierite (FER) and smectite (SM) adjacent to the glass. Micro-analysis and morphology confirmed the identify of ferrierite.

Fig. 9. Backscattered electron image ofcross section ofdenselyintergrown ferrierite and orthoclase (ferrierite tuffSR67), the lat-ter showing the relict shard position.

Fig. 10. Fracture section of sample SR67 showing ferrieritehabits. (a) Radiating bundle of fibers. (b) Boxy acicular crystals.

alteration of the unstable natural glasses. Hay (1963)showed that the early-formed clay in the John Day For-mation, resulting from percolating ground waters, was animportant agent in altering the water chemistry, render-ing it favorable for zeolite crystallization by raising thepH.

There are several reasons for suggesting that ferrieriteis probably the first zeolite to appear in the transforma-tion ofglass: (1) ferrierite occurs on silica and clay sub-strates in core 2 tuffs, in combination with orthoclase andsmaller amounts of clinoptilolite, (2) ferrierite coexistswith glass and clay in core 4 tuffs, and (3) the verticalzonation within the deposit as a whole shows a progres-sion from glass to ferrierite to mordenite.

At Lovelock, cristobalite and clinoptilolite cocrystal-lized as replacements for dissolving glass, probably afterthe onset of the crystallization of ferrierite, mordenite,and orthoclase. The occasional draping of mordenite onclinoptilolite (SRl.5) suggests thaI, at least in those par-ticular tuffs, mordenite postdated the clinoptilolite.

Based on their coexistence in many tuffs, mordenitegrowth occurred during or after ferrierite crystallization.

RICE ET AL.: FERRIERITE CRYSTALLIZATION J Z I

Fig. I l. SIMS images of a cross section of sample SR67. Most of the areas of high K concentration represent orthoclase. Thetwo grains in the bottom ofthe field are sanidine pyroclastic grains.

There is no convincing evidence that one crystallized atthe expense of the other. Both ferrierite and mordenitecrystallize into open pore space in the tuffs, but there wasno evidence from TEM characterization that these twozeolites crystallized in close proximity in the fashionthat ferrierite and orthoclase did. The textural preferenceof mordenite for filling pores produced by elimination ofshard grains, requiring as it does an advanced stage ofglass dissolution, is interpreted to occur later than theintimate growth of ferrierite and orthoclase.

Cristobalite is a minor to major phase in the hole no.2 tuffs as well as in some ferrierite tuffs. It also coexistswith clinoptilolite and orthoclase, but there was no pet-rographic evidence of the time relationship of cristobaliteand ferrierite.

Orthoclase appears to have been the last phase to crys-tallize. The orthoclase textures in the ferrierite tuffs, suchas its presence at the periphery of former shards or con-

centric to adjacent shard pore outlines, could indicatecrystallization contemporaneous with ferrierite. In theseinstances, orthoclase crystallization occurred along zonesof microchemical enrichment of K and Al. Overgrowthsand crack-fillings of orthoclase-rich material suggest thatin those instances it persisted after the main episode ofzeolitrzahon.

Figure l3 schematically summarizes the parageneticsequence ofauthigenic phases in the Lovelock deposit.

Bur,x crrynrICAL AND MTNERALOGICAL TRENDS

Bulk chemistry of core hole no. 3

Bulk chemical data for core samples from hole no. 3

fig. 1a) suggest that several distinct units are representedin the 30-m thick deposit of rhyolite pyroclastics. Al-though the average pyroclast grain size, as inferred fromthe textures in SEM images, seems to be relatively con-

322 RICE ET AL.: FERRIERITE CRYSTALLIZATION

Tlele 4. Mordenite (1 , 2) and clinoptilolite (3, 4) chemical anal-yses

sio,Alro3MgoCaONaroKrOTio,FeO

Total

Remarks

oo.o10.20.002.0s0.240.840.010.09

80.02

vug filling

68.310.80.002.170.971 . 1 00.000.08

83.32

vug filling

72.813.21 . 1 64.000.450.960.000.00

92.59

rhombus

76.813 .50.844.710.200.780.000.00

96.87

rhombus

Fig. 12. Cross sections of mordenite tuff SR3 showing (a)individual void space, outlined by orthoclase crystallite networkand partially filled with prismatic mordenite, (b) double-walledtexture oforthoclase and minor cristobalite.

stant over the entire length of core 3, the chemical vari-ations require the recognition of several distinct units.Phase identification based on PXD helped in delineatingthe zones. From the surface to about l0 m is a ferrierite-rich zeolitic zone (layer A), containing some authigenicorthoclase and cristobalite but whose mineral content ap-proaches pure ferrierite in some places. These tuffs appearto have the same mineralogical characteristics as the fer-rierite-rich rocks that crop out in the southwest corner ofthe area. Between l0 and 12.5 m is an Mg-rich, low Si/Al zone (layer B) containing ferrierite and smectite. Atabout the 12.5-m level there is a break in the dominantassemblage. Below 12.5 m ferrierite is absent and, withthe exception of some clay-rich beds and pyroclastic frag-ments, only vitric ash (layer C) is present. Judging fromthe material recovered at 30 m, smectite rims probablyoccur on most of the glass shards below 12.5 m, even inthose beds that are primarily vitric. This indicates that,although smectite alteration preceded ferrierite forma-tion, not all of the clay-altered tuffs were zeolitized. The

composition of this clay is Mg rich, as distinct from laterclays, which were detennined by TEM/EDX to be Ferich. These later clays occur as thin alteration productson ferrierite laths and are not considered important inthe context of ferrierite genesis. As shown in Table 5,layers A and B both have higher MgO than does layer C.Because the initial bulk chemical compositions of A, B,and C were probably comparable, zeolite-forming reac-tions must have involved some additional chemicaltransformation specific to the tufs shallower than 12.5 m.

Bulk compositions of the upper zeolitic tuffs in holeno. 3 show decreasing alkali-metal concentrations withdepth (Fig. l4). In the near-surface tuffs of layer A theKrO concentration (6-80/o) is higher than in fresh rhyolite,but KrO gradually decreases to a relatively low value inlayer B. Na content also decreases downward over thisinterval, whereas Ca increases gradually. The SEM datasupport the presence of K-enriched zones within the layerA tuffs, and it may be that there is more authigenic feld-spar in the near-surface tuffs. Many examples of KrO en-richment, such as authigenic orthoclase overgrowths onrelict pyroclastic feldspars and feldspar-rich veinlets inthe zeolitic tuffs, reflect the presence of a K-rich fluid.The question remains whether the K and Na were addedfrom an external source or reflect variations in originaltuffchemistry. Fractionation of KrO is not uncommon insome ignimbrite flow sequences, such as the Bishop Tuff(Hildreth, 1979) and the Valley of Ten Thousand Smokes(Mahood, 1986), in which increasingly less silicic lavasextrude onto earlier more silicic flows. However, if thiswere the case, K2O and Si/Al values from the base oflayer B to the top of layer A in hole no. 3 would beexpected to decrease. They demonstrate precisely the op-posite trend. It also seems unlikely that concentrationsvarying by a factor of 4 over just l5 m would result fromcontinuous changes in the source chemistry or that anair-fall pyroclastic deposit would reliably record thesechanges. A more probable explanation is the addition ofalkali-rich fluids or K-rich fluids that partially exchangedwith Na in the deposit, thus reducing the total Na con-centrations below typical values for rhyolites.

Mass balance

In order to understand the chemistry of zeolitization,particularly with respect to ferrierite crystallization, iI is

RICE ET AL.: FERRIERITE CRYSTALLIZATION 323

magnesrumsmectite l - . " "" -_ l i ronsmect i te l ; - -T

';:,:"w

cristobalite

ctinoptitotite W

rordeni teW

opaline veins I I

Time _-_____________>

Fig. 13. Paragenetic sequence ofauthigenic minerals recog-nized in the Lovelock deposit.

necessary to determine whether the reactions were iso-chemical or required significant fluxes of material. Changesin elemental concentrations were calculated assuming thatthe composition of core 3 vitric tuffs of layer C was rep-resentative of the original tuffs from which the zeolitesformed. Layers A and B were considered individually. Itis not known a priori that any of the components weretruly immobile or insoluble during the hydration reac-tions. The compositions were calculated on an H.O-freebasis in order to determine the relative change in A andB from C. Rhyolites or obsidians usually contain only afew percent HrO, whereas ferrierite tuffs have approxi-mately 100/o HrO. It was also assumed consistent withmetasomatism that the volume of the pyroclastic tufsremained approximately constant during the process. Thisis the metasomatic assumption. The following relation-ship was used to calculate A values:

Ao*iae : 100(oxide".po - oxide..p./oxide..p.) (l)

in which oxideo is the weight percent of the oxide in layerA and po is the density of tuff layer A. The equivalentequation for layer B relative to layer C was also applied.

There are at least two possibilities that explain the dif-ferences between layers A, B, and C: (l) that they wereseparate and chemically different units to begin with, or(2) that concentration changes were produced by the ad-dition or subtraction of material. Both may be correct.Layers A and B have lower Na contents than does C,which suggests leaching. The fate of the Na is not clear,however. Because the K concentration tends to be rela-tively high and that of the Na low, an alternative mech-anism is coupling through exchange reactions. The ob-servations suggest that the reactions assisting zeolitecrystallization were not isochemical. All three layers havemuch higher Mg, higher Ca, and lower Na concentrationsthan would be expected for a rhyolite. The large relativedifferences determined for A and B with respect to Csuggest also that these were deposited separately. The Al-,Fe-, and Mg-rich composition of layer B suggests that itoriginally had a larger mafic component than did layer A.

Because ferrierite is a high-silicazeolite, a high activityof silica in the zeolite-forming fluids is required. Many of

t 2 1 6 b

Al2o3

0 1 2 3 { 5 6 0 r 2 3 0 1 2 3 r 0 2 4 6 a

MgO CaO Na2O KzO

Fig. 14. Variation of element concentrations with depth inhole no. 3 tufs. Layer A is predominantly ferrierite and ortho-clase. Layer B is predominantly ferrierite and smectite. l,ayer Cis glass and clay.

the vug occurrences in andesite and basalt have liningsof chalcedony in intimate contact with ferrierite (Gottardiand Galli, 1985). The presence of cristobalite implies thepresence offree silica during the zeolite diagenesis in theLovelock deposit. Abundant opaline rocks visible in out-crop are further evidence.

The total Mg content of the original rhyolite, only about0.50/o MgO, was insufficient to account for the vast quan-tities of ferrierite in the Lovelock deposit. The additionof Mg in the form of smectite coatings on glass shardsappears to be responsible for most of the Mg necessaryfor ferrierite formation.

Ferrierite crystal chemistry

Several populations of ferrierite samples were identi-fied by electron microprobe analysis, namely, a group withMg/Ca of 2.5 to 3.0 and a group with Mg/Ca of about0.5. Both compositional groups are represented in KP52(e.g., Table 2, analyses I and 2) and SR64 tuffs. DuringTEM/EDX work, substantial variation in Si, Mg, and Caconcentrations was observed for crystals from tuff speci-mens as small as I mm3 (Fig. l5). The grouping of fer-rierite compositions into high and low Mg/Ca groups ap-

Tmle 5. Mass balance in hole no. 3 ferrierite tuffs.

Upper Lower(A) (B)

ferrierite ferrierite Vitric tuffzone zone hole 3

oxide (7)* (4)-- (6)-. a (A), % a (B), %

sio,Al2o3Na.OKrOMgoCaOFe,O"Tio,

Total

77.8 71.4610.3 16.491.50 0.785.39 3.741.06 3.571.12 1 .981.32 2.750.15 naf

98.64 98.02

-2 .O - 10 .0-23 +22-45 -72-4.5 -34

+59 +435-34 - +16-0.30 +45

0 naf

75.012.732.595.390.631 .611 .790 .15

99.83

'Mass changes for layers A and B assuming C is the original com-position (dry basis) using Equation 1 with bulk compositions of core 3 tuffs.

" Number of samples averaged in parentheses.t The abbreviation na : not analyzed.

324 RICE ET AL.: FERRIERITE CRYSTALLIZATION

1 . 5

o M g

o c a 1 0

Lovelock ferrierite (KP52)

(J

nv

a ) na ) v

( t v

LOVELOCK PHASE COMPOSITIONS

Fresh \ - r l

Na2O + K2O CaO

Fig. 16. NormalizedMgO-(NarO + KrO)-CaOcontents(wtYo)for Lovelock zeolites and tuffs. The dashed field represents about40 bulk analyses of tuffs from Lovelock.

indicate a fundamental role of Mg-bearing fluids in theformation of ferrierite in the Lovelock deposit.

DrscussroN

Comparison of Lovelock tuffs with other silicic pyroclastictuffs

Of the many potential controls on ultimate zeolite as-semblages, one that is constrained at Lovelock is the bulkglass starting composition, represented by the fresh glassanalyses of C3-30.5, SR78, and C4 tuffs. The chemicalcompositions representative of the three layers A, B, andC in core 3 are given in Table 5. In Figures l7a-17d,Lovelock compositions are compared with those of ap-proximately 40 other pyroclastic glassy rhyolites from theliterature. It was desirable to have the most direct com-parison with like tuffs, namely, pyroclastic fall depositsof silicic composition, rather than the whole rhyolitic suite.Some of these have zeolites associated with them, butmost do not. Chemical data were obtained for Lovelockglass compositions by electron microprobe analysis, SIMSimaging, and TEM/EDX. Lovelock bulk ash (layer C) hasapproximately two times more MgO and about one andone-half times more CaO than other pyroclastic rhyoliteswith the same SiO, content (Figs. l7a, l7b). The MgO(<0.05 wto/o) and CaO (-0.2 wto/o) concentrations deter-mined for the interior of the Lovelock shards were righton the trend for a silica value of 780/0.

When the compositions of layers A and B are com-pared with the silicic trend, it is found that MgO andCaO contents are, in general, much higher than expectedfor silicic glassy tuffof the same silica content. K concen-trations are typical for rhyolites, whereas Na concentra-tions are about a factor of 2 below normal. Thus, al-though layer C has apparently not been significantlyleached ofNa, layers B and C probably have been, sug-

0 0 i -30

trt r oSi/unit cell

Fig. 15. Divalent cation concentrations vs. Si for individualferrierite crystals from a small chip of ferrierite tuff(KP52).Datawere obtained from TEM/EDX analyses, based on unit-cell con-tents of 72 O atoms.

pears to be a consequence ofCa concentration differences.Another grouping of ferrierite compositions occurs in thecore 2 tuffs but in smaller amounts than in core 3 tuffs.The cluster of three calcic ferrierite compositions in Fig-ure 16, outside the compositional field defined by bulkzeolitic tufl probably reflects in part a chemistry inher-ited from the original core 2 tuff.

Compositions of glass, ferrierite, and smectite dem-onstrate the chemical control of the precursor glass andadded smectite on the resultant ferrierite (Fig. 16). Theferrierite compositions (solid circles) reside between thealkali-rich glass and the (Mg + Ca)-rich smectite (opencircle). Two inferences can be drawn from this. First, theferrierite in the Lovelock deposit, although alkali-richcompared with many other ferrierite compositions, hastoo much Mg to be derived in large quantity from thepristine glass and HrO alone. All known natural ferrieritesamples have at least 0.50/o MgO. Second, the mordenitecompositions (solid squares in Fig. l6) and the clinoptilo-lite compositions (open squares) are too calcic to be de-rived from the glass + HrO + smectite. This is one keyto the subordination of mordenite and clinoptilolite rel-ative to ferrierite in this locality because it is not likelythat either mordenite or clinoptilolite could accept suchlarge amounts of Mg and remain stable.

To sum up: Mg occurs at significant concentrations inall of the known natural ferrierite occurrences. The Mgconcentration per unit cell, based on 72 O atoms, rangesfrom approximately 0.4 to nearly 3, averaging around 2(Sameshima, 1986). The Lovelock ferrierite, although ithas a small Mg content compared with many others, nev-ertheless required the addition of substantial Mg to thesilicic glass for development in large quantities. Mg hasa specific site in the ferrierite structure, for which it isespecially suited, and it is not easily removed by acidleaching (Vaughan, 1966). These observations strongly

RICE ET AL.: FERRIERITE CRYSTALLIZATION 325

z0

1.5

3.s7t oB

OA

.... t lo oc

'' lj'i... co

Mgo

gesting that in addition to the clay formation, other hy-dration reactions accompanied zeolitization.

Geological setting

Of the many sources of aqueous fluids in which zeolitesoccur, those possibly important for the Lovelock depositare hydrothermal, percolating ground water, and lacus-trine. There is an absence of minerals diagnostic of hy-drothermal activity. The widespread silicification doeshave the appearance of a fused material, but this featureneed not be a high-temperature reaction because the silicacould be derived from feldspar-forming reactions at lowtemperature. The authigenic feldspar itself, authigenic or-thoclase, although perhaps representative of the highesttemperature or highest grade of alteration attained atLovelock, need not be ofhydrothermal origin. For thesereasons, a hydrothermal origin seems unlikely.

CaO

Lacustrine deposition could well be involved, as evi-denced by the presence of diatomite beds, authigenic feld-spars that commonly occur in saline, alkaline lake de-posits (Hay, 1966), and the relatively sharp boundarybetween the zeolite tuffs and the underlying glassy beds.From SEM observations of morphological characteris-tics, the diatoms were identifiedas Melosirq crenulata orMelosira granulata. As far as the authors are aware, thesespecies are not constrained by salinity. If connate lakewater is invoked as the sole agent of alteration, the pH

of the water could not have exceeded about 9 withoutdissolving much of the opaline silica. The fairly sharpcontact between vitric tuff and ferrierite tuff in hole no'3 could be interpreted several ways. One interpretationwould favor a lacustrine depositional environment be-cause the bottom ofthe lake would be represented by anabrupt break. A fault juxtaposing relatively fresh tuffs

1.0

0.0

SiO272 74

SiO2

Na2O Kzo

SiO270 72 74

SiOzFig. 17. Diagrams of variations in element concentrations for bulk analyses from silicic pyroclastic tuffs. Data taken from

literature sources are solid circles; Lovelock layers A, B, and C, as well as the fresh Lovelock glass (G) in layer C, are represented

by large open circles.

807A766880787674

807A687674727068

a

.Boa

3.'a ao

aoa3

ad

.C oo aI' r ' r .

o a o na ^\:

OA

. . o3

a

.oG

OA

o

O 3

O

a

326 RICE ET AL.: FERRIERITE CRYSTALLIZATION

with highly zeolitized, tuffs could also explain the discon-tinuity. The vertical changes in alkali concentrations andthe presence ofdiatomaceous beds could both be due tothe presence of a relatively freshwater lake. However, theresults of mass-balance calculations described above,which show addition of both Mg and K and the removalof Na, strongly argue for an open ground water system.

Chemical and textural development during diagenesisBecause of the widespread occurrence of Mg-rich clay

coatings on the glass shards and the relatively minor con-centrations of alkaline-earth cations in the fresh glass, itis inferred that Mg2* and Ca2* were added to the vitricpile by aqueous fluids sometime during or after deposi-tion. This could have been accomplished through eitherthe presence of brackish connate HrO derived from asaline lake or by leaching ofnearby basalts and transportof Mg and Ca to the Lovelock site. The smectite also hadimportant implications for the textural development ofthe deposit. The Lovelock zeolite tuffs, as do most otherzeolite tuffs, preserved a vitroclastic texture inherited fromthe precursor ash. Textural observations made in thisstudy point to early smectite as a key factor in this pro-cess. The presence of smectite at the peripheries of shardgrains was responsible for preserving the vitroclastic tex-ture, whereas the porous nature of the coating permittedglass dissolution to proceed. Stanley and Benson (1979)observed hollow clay coatings of smectite cements in vit-ric rocks from which detrital grains had been dissolved.The observed variations in Mg content of the Lovelockferrierite samples noted earlier could be caused by theproximity of the crystallizing ferrierite to smectite or byvariation in smectite composition.

The widespread occurrence ofauthigenic orthoclase atLovelock is a feature that was not fully recognized inprevious descriptions ofthe deposit. According to a studyby Sheppard and Gude (1973), zonation of minerals oftenoccurs in lacustrine zeolite deposits where saline mineralsare interbedded with zeolites. These typically exhibit asequence proceeding from glass under freshwater condi-tions to zeolites under moderately saline conditions topotassium feldspar and searlesite under highly saline con-ditions. No saline minerals have been recognized in thedeposit. Thus, rather than indicating the extent of dia-genesis in a highly saline environment, crystallization oforthoclase in the Lovelock deposit may be more a con-sequence ofhigh K activity in the altering fluids.

Cristobalite and clinoptilolite formed as diagenetic re-placements of glass shards. One of the petrographic fea-tures observed (e.g., C2-17.3) was relict shards replacedby silica and discrete crystals of clinoptilolite. This re-placement was probably the result of a dissolution-pre-cipitation mechanism in which local shards were dis-solved, with or without the presence ofa gel phase, andreprecipitated within the same local volume. This processcould have occurred on a shard-sized scale because thedissolution of a rhyolite glass shard will yield a compo-sition similar to silica and any of the zeolites represented.

Iijima (1971) noted that clinoptilolite analyses deter-mined by electron microprobe were anomalously high insilica and attributed this to finely divided silica mixedwith the zeolite. Several similar observations were foundin the course of the present work as well.

Na was either leached from the beds or exchanged withK prior Io zeolitization. Perlite textures point to the like-lihood of Na leaching, since Na loss usually accompaniesthe early stages ofglass hydration (Jezek and Noble, I 978).A transfer of Na could have had an important influenceon the resultant assemblage because ferrierite, althoughalmost invariably containing minor Na, does not preferit as mordenite does. Furthermore, although reactions inthe presence of high c"" could have precipitated phasessuch as albite or analcime, such relations are not foundat Lovelock. Presumably because the prevailing silicaconcentrations were high, mordenite is the main Na-bearing phase. It is possible that low Na concentrationswere necessary in addition to high Mg concentration inorder that large quantities of ferrierite could crystallize.

The onset of ferrierite crystallization did not requirethe complete dissolution of the shards because it initiallyformed within primary pores and utilized glass and claysas substrates. The coarse tuffsample 4-8.5 is an examplein which a mostly vitric tuff shows incipient ferrieritecrystallization (Fig. 7). Here, where pore space was am-ple, the prismatic habit developed. The ferrierite-rich tuffSR67 provides an example of a case where the space wasmore limited or in which the activity of K was sufficientlyhigh. As a consequence, finer, more massive ferrierite co-crystallized with orthoclase (Fig. 8). In this case, the habitof ferrierite is not obvious. Figure 8 illustrates the chem-ical heterogeneity within microscopic zones in many ofthe tuffs. The clinoptilolite and cristobalite replaced theglass while the K-rich phases ferrierite and orthoclase to-gether fill the primary pore space.

Ferrierite is a rarely occurring zeolite and is unusualfor its Mg-rich composition. Studies of the known naturalferrierite compositions (Wise and Tschernich, 1976; Sa-meshima, 1986) reveal that, without exception, ferrieriteis a high-silica zeolite with a narrow range of frameworkSi/Al ratios (-4.5-6) and that Mg is an essential ingre-dient. The conditions for ferrierite crystallization in theLovelock deposit appear to require both high concentra-tions of silica, supplied by the rhyolite, and of Mg. Haw-kins and Roy (1963) pointed out that clay minerals willnormally crystallize from Mg-bearing solutions formedby glass dissolution. The fact that ferrierite is so abundantin the Lovelock deposit must depend not only on suchinfluences as pH, temperature, and composition but alsoon the nature of structural units present in the fluids.Certainly the combination of high Mg and high Si wasimportant for ferrierite to crystallize. It is also possiblethat a templating process analogous to that proposed forlarge organic cations in some synthetic zeolites (Barrer,l98l) was operative during diagenesis in the Lovelockdeposit. As noted above, Mg has a special role in theferrierite structure. The presence of hydrated Mg ions,

RICE ET AL.: FERRIERITE CRYSTALLIZATION 32'l

Mg(HrO)l+, in crystallizing fluids likely fostered the knit-ting together of the ferrierite tetrahedral framework. Be-cause the K ions can reside indiscriminately in the fer-rierite channels, responding to framework charge but notaffected by stringent steric considerations, the K concen-trations are probably an accident of the prevailing solu-tion chemistry.

The differences in ferrierite composition from one bedto another are indicative of the local chemical control onferrierite formation and the lack of homogenization ofdiagenetic fluids on a large scale. Which of the zeolitescrystallized was a function of the chemical compositionof the mineralizing fluid, and this, in turn, was controlledby the compositions of parent glass and the mineral frag-ments in the pyroclastic deposit. The ferrierite chemicaldata obtained by TEM/EDX also suggest (e.g., Fie. 15)that the aqueous chemistry can vary on a very fine scale,perhaps approaching the volume of a few pores. We knowthat pH can have alarge effect on the proportions ofAland Si species in aqueous systems. For example, the ratioof two critical ionic species, Si(OH)"/Al(OHh, may varyby as much as a factor of 4 between pH 9 and l0 (Marinerand Surdam, 1970). This diagenetic system is clearly oneof several scales of mass transfer through time. The firstevent, magnesium smectite formation, was the result ofwidespread but heterogeneous influx of fluids into thetuffs. Crystallizarion of zeolites was apparently controlledon a much smaller scale.

AcrNowr,nocMENTS

This work represents a portion of the first author's Ph.D. dissertation

at Lehigh University. He wishes to thank Exxon Research and Engineer-

ing Co. for supporting this research and allowing its publication, and

M.M.J. Treacy and A.J. Jacobson, in particular, for their encouragement.He is also grateful to Charles B. Sclar for his helpfulness as the major

advisor. SIMS microscopy was provided by D.P. kta and W.B. Lamberti

of Exxon Research and Engineering Co. and is geatly appreciated. Re-

views by R.J. Iander and an anonymous reviewer and editorial commentsby D.L. Bish contributed to significant improvements in the manuscnpt.

RrrnnsNcrs cITED

Barrer, R.M. (1981) Z€olites and their synthesis. Zeolites, l, 130-140.Boles, J.R. (1972) Composition, optical properties, cell dimensions, and

thermal stability of some heulandite group zeolites. American Miner-alogist, 56, 1463-1493.

Boles, J.R., and Coombs, D S. (1975) Mineral reactions in zeolitic Triassictuff, Hokonui Hills, New Zealand. Geological Society of America Bul-letin, 86, 163-173.

Boles, J.R., and Surdam, R.C. (1979) Diagenesis of volcanogenic sedi-ments in a Tertiary saline lake; Wagon Bed formation, Wyoming'American Journal of Science,279, 832-853.

Brey, G., and Schmincke, H.-U. (1980) Origin and diagenesis of the RoqueNublo breccia, Gran Canaria (Canary Islands)-petrology of RoqueNublo volcanics, II. Bulletin ofVolcanology, 43, 15-33.

Broxton, D.E., Bish, D.L., and Warren, R.G. (1987) Distribution andchemistry of diagenetic minerals at Yucca Mountain, Nye County, Ne-vada. Clays and Clay Minerals, 35, 89-l 10.

Cormier, W.E., and Sand, L.B. (1976) Synthesis and metastable phase

transformations of Na-, Na,K-, and K-ferrierites. American Mineral-ogist,6l, 1259-1266.

Davies, D.K., Almon, W.R., Bonis, S.B., and Hunter, B.E. (1979) De-position and diagenesis of Tertiary-Holocene volcaniclastics, Guate-

mala. Society of Economic Paleontologists and Mineralogists Special

Publication, 26, 281-306.Goldstein, J.I., Costley, J.L., Lorimer, G.W., and Reed' S'J'B' (1977)

Quantitative X-ray analysis in the electron microscope- Scanning Elec-

tron Microscopy/ 1977, 1, 315-323.Gottardi, G., and Galli, E. (1985) Natural zeolites. Minerals and Rocks

18, 409 p. Springer-Verlag, Berlin.

Graham, R.P.D. (1918) On ferrierite, a new zeolitic mineral from British

Columbia; with notes on some other Canadian minerals. Transactions

of the Royal Society of Canada, 3 (12)' 1 85-20 I

Hawkins, D B, and Roy, R. (1963) Distribution of trace elements be-

tween clays and zeolites formed by hydrothermal alteration ofsynthetic

basalts. Geochimica et Cosmochimica Acta, 27 ' 785-795.

Hay, R.L. (1963) Stratigraphy and zeolitic diagenesis of the John Day

Formation of Oregon. University of California Publications in Geolog-

ical Sciences, 42, 199-262.-(1966) Zeolites and zeolitic reactions in sedimentary rocks. Geo-

logical Society ofAmerica Special Paper, 85, 130 p.

Heiken. G.. and Wohletz, K. (1985) Volcanic ash,246 p' University of

California Press, Berkeley, California.Hildreth, W. (1979) The Bishop tuff Evidence for the origin of compo-

sitional zonation in magma chambers. Geological Society of America

Special Paper, 180, 43-75'Iijima, A. (1971) Composition and origin of clinoptilolite in the Naka-

nosawa tuffof Rumoi, Hokkaido. In Molecular Sieve Zeolites-I, Ad-

vances in Chemistry Series, l0l, 334-341.

Jezek, P A., and Noble, D.C. (1978) Natural hydration and ion exchange

of obsidian: An electron microprobe study. American Mineralogist, 63,

266-273.Jones, J.B., and Segnit, E.R. (1971) The nature ofopal I' Nomenclature

and constituent phases. Joumal of the Geological Society of Australia,

l8 ( l ) , 57-68Keller. W.D., Hanson, R.F., Huang, W.H.' and Cervantes, A' (1971) Se-

quential active alteration of rhyolitic volcanic rock to endellite and a

precursor phase of it at a spring in Michoacan, Mexico. Clays and Clay

Minerals, 19, l2l-127.kta, D.P. (1935) A high-resolution, single ion sensitivity video system

for secondary ion microscopy In A. Benninghoven, R.J' Colton, D'S

Simons, and H.W. Werner, Eds., Secondary ion mass spectrometry

(SIMS v), Springer Series in Chemical Physics, vol. 44' p' 232-234'

Springer-Verlag. New YorkMahood, G. (1986) Waxing and waning magma chambers: A comparison

of zoning in ignimbrites and plutons. Fourteenth International Min-

eralogical Association Abstracts with Program, 163.

Mariner, R.H., and Surdam, R.C. (1970) Alkalinity and formation of

zeolites in saline alkaline lakes. Science, l7O'977-980-

Mason. B , and Sand, L.B. (1960) Clinoptilolite from Patagonia' The re-

lationship between clinoptilolite and heulandite. American Mineralo-

gist ,45, 341-350.

Mumpton, F.A. (l 960) Clinoptilolite redefi ned. American Mineralogist,

45,35r-369.

Noh. J.H., and Kim, S'J. (1986) Zeolites from Tertiary tuffaceous rocks

in Yeongil area, Korea. In Y. Murakami, A' Iijima, and J'W' Ward'

Eds., New developments in zeolite science and technology' 28, p' 59-

66. Proceedings ofthe Seventh International Zeolite Conference, Ko-

dansha and Elsevier, TokYo.

Regis, A.J. (1970) Occurrences offerrierite in altered pyroclastics in cen-

tral Nevada. Geological Society of America Abstracts with Program,

2, 661.Sameshima, T. (1986) Ferrierite from Tapu, Coromandel Peninsula, New

Z,ealand, and a crystal chemical study of known occurrences' Miner-

alogical Magazine, 50, 63-68.

Sand, L.B., and Regis, A.J. (1968) Ferrierite, Penhing County, Nevada.

In Abstracts for 1966, Geological Society ofAmerica Special Paper,

l 0 l , 1 8 9 .

Sheppard, R.A., and Gude, A.J., III (1968) Distribution and genesis of

authig,enic silicate minerats in tuffs of Pleistocene l-ake Tecopa, Inyo

County, California. U.S. Geological Survey Professional Paper,597,

38 p.-(1973\Zqolites and associated authigenic silicate minerals in tuff-

328 RICE ET AL.: FERRIERITE CRYSTALLIZATION

aceous rocks of the Big Sandy formation, Mohave County, Arizona. Stewart, J.H., and Carlson, J.E. (1978) Geologic map of Nevada. U.S.U.S. Geological Survey Professional Paper, 830, 36 p. Geological Survey, scale 1:500000.

Sheppard, R.A., Gude, A.J., III, and Mumpton, F.A. (1983) Field trip Vaughan, P.A. (l%6) The crystal structure of the zeolite ferrierite. Actastop 6, Invelock zeolite deposit, Lovelock, Nevada. In F.A. Mumpton, Crystallographica,2l,983-99O.Ed., International Committee on Natural Zeolites Zeo-Trip '83, p. 48- Wise, W.S., and Tschemich, R.W. (1976) Chemical composition of fer-55. Allen Press, Larvrence, Kansas. rierite. Arnerican Mineralogist, 6I, 60-66.

Stanley, K.O., and Benson, L.Y. (197q Early diagenesis of high plainsTertiary vitric and arkosic sanilstone, Wyoming and Nebraska. SocietyofEconomic Paleontoldgists and Mineralogists Special Publication,26, Maruscrrrr RrcErvED FesnuA,ny 22, l99l401-423- Mercuscnrpr AccEprED Noverrlsen l. l99l