Amefican Mineralogist, Volwe 70, pages 499-516, 1985

Metasomatism, titanian acmite, end alkali amphiboles in lithic-wacke inclusionswithin the Coyote Peak diatreme, Humboldt County, California

GsRAro K. Cz,c,MANsKE AND SrevEN A. ArrrNr

U.S. Geological Suruey345 Miilillefield Roail, Menlo Park, California 94025

Absfirct

Lithic-wacke inclusions within the alkalic ultramafic diatreme at Coyote Peak record aremarkable history of metasomatism and crystal growth. Dominant metasomatic changeswere loss of Si from the inclusions and mass influx of K, due to an unusually high K activityin the ultramafic host. Reaction with K converted much of the clay, quartz, and lithic fractionof the lithic wacke into microcline and exchanged Na from abundant clastic albite.

Sodium released from albite: (1) combined with Ti and Fe to form myriads of acrnitecystals, many of which are strongly zoned with cores unusually rich in titanium; (2) com-bined with Ti, Fe, Mg, and Ca to form alkali amphiboles, which are zoned from fluor-richterite cores to unusual titanian-arfvedsonite rims; and (3) formed a late-stage zeolitesimilar to natrolite. The sequence of ferromagnesian phases is interpreted to have beencontrolled partly by falling temperature, but dominately by falling /o, as the oxidized sedi-mentary assemblage equilibrated with the reduced ultramafic melt.

Preservation of primary sedimentary textures to within a few millimeters of the lithic-wacke/ultramafic interface in zoned inclusions and strong zonation between the cores andrims of individual phases suggest that recrystallization and chemical exchange were of shortduration because of rapid emplacement and cooling of the diatreme at shallow depth.

Introduction

An alkalic, ultramafic diatreme 260 by 500 m in outcroparea penetrates a lithic-wacke sandstone sequence of theFranciscan assemblage, 20 km SE of Orick, at CoyotePeak, Humboldt County, California. Rocks comprising thediatreme may be conveniently subdivided into uncon-taminated and contaminated variants, based on SiO, con-tent, color, mineralogy, the character of contained clots,and abundance of inclusions. Poor exposures do notpermit these variants to be mapped separately. Uncon-taminated rocks comprise two distinct assemblages, eachcontaining phenocrysts of subhedral to rounded olivine (to2 mm; reverse zoned from Foro to Forr; NiO:0.M,MnO:0.4 and CaO:0.5 wt.%) and anhedral t itano-magnetite (zoned from 13 to 9 wt.oh TiO, and from 5 to 4wt.% MgO). Groundmass phases in the most pristine as-semblage are phlogopite (xonrog : 0.8 to 0.9), melilite[(CarFe)rr(CarMg)57(CaNaAl)r r], nepheline (NetrKs2r),magnetite (TiOz < 7 and MgO < 2 wt.o/o), perovskite, apa-tite, sodalite, and poikilitic garnet containing up to 16wt.%o TiOr. A second assemblage, characterized by coronasof pyroxene plus phlogopite around olivine, contains essen-tial augite (WoroEnn.Fst) and nepheline (Ne^Ks2r) andlacks melilite. Study of many thin sections shows the as-semblages to be related to one another by an evolutionary,

r Present address: Molycorp, Inc., Questa Division, P. O. Box469, Questa, New Mexico 87556.

0003{xxx/8 5/050H499$02.00

reaction process. Rocks characterized by both mineral as-semblages are black; they also contain discrete coarse-grained clots (to 6 cm) composed of phlogoPite, garnet(zoned from 12 to 2 wt.% TiO2), minor sodalite and pec-tolite, pyrrhotite, and at least five K-Fe and Na-Fe sulfideminerals (Czamanske et al., 1979,1980, and 1981; Erd andCzamanske, 1983). Considering mineral assemblages, pris-tine rock at Coyote Peak might best be termed modlibovite(Scheumann, l9l3). Contaminated rocks are greenish incolor, but essentially are composed of the reacted mineralassemblage found in the uncontaminated rocks. Also pres-ent are pectolite and natrolite, and a feathery garnet that isless Ti-rich [8 to 1 wt.%) and more aluminous (2.6 to 6wt.%) than garnet in uncontaminated rocks. Contaminatedrocks contain a totally different type ofzeolitic clot, rich innatrolite and containing acmite, alkali amphibole, biotite(Xpnro, to 0.,14), titanite, and barytolamprophyllite (8'E to26wt.%BaO).

The contaminated rocks contain reacted fragments ofmudstone, as well as rounded inclusions of lithic wackepicked up from the Franciscan assemblage at the time ofemplacement. These lithic-wacke fragments have under-gone substantial chemical exchange with the diatreme, in-cluding extreme potassic metasomatism. Small (less than3-cm diameter) inclusions, which are uniformly pale graygreen, initially were thought to represent fine-grained po-tassic igneous rocks (Czamanske et al., 19771. Larger (diam-eter 6-10 on) fragments commonly have brown cores anddistinct gray-green rims, 8 to 12 mm thick, that closely

499

500

Fig. 1. Specimen from the Coyote Peak diatreme containingfour lithic-wacke inclusions- Inclusions on left and upper right aretypical of the more uniformly altered, small inclusions (e.g., sampleCYPS); the two larger, rounded inclusions show distinct core andflm components.

resemble in color and texture the smaller, more uniformlyaltered inclusions (Fig. 1).

Due to the extreme disparity in composition between theultramafic host and the included lithic-wacke fragments(Table l), this study provides a special opportunity to de-scribe pronounced metasomatic exchange. Resulting min-eral assemblages contain titanian acmite and alkali am-phiboles of unusual interest.

PetrographyWe selected eight samples for study, including: (l) uni-

formly altered small inclusions (CYP8 and 103); (2) amedium-size inclusion (CYP52); (3) large inclusions withdistinct core and rim alteration facies (Cyp7, 101, 102, and178); and (4) a sample of lithic-wacke country rock (Cyp4)collected from a surface outcrop a few meters beyond thediatreme contact (which is nowhere exposed). In addition,sample CYP5 was selected as representative of the uncon-taminated ultramafic melt.

The country rock (CYP4) is a light-tan well-induratodmassive fine- to medium-grained moderately sorted lithicwacke. By volume it is composed of quartz (20%); albite(10%; Ab"" u); lithic fragments, including shale, chert, andminor volcanic- and metamorphic-rock fragments (40%);and a turbid clay-rich matrix (25o/o). Altered biotite, chlor-ite, and muscovite (total, approximately 5oh) arc also pres-ent (Table 6). Franciscan graywacke of the northern CoastRanges of California is noted for the absence of potassiumfeldspar (Bailey et al.,1964), and no potassium feldspar wasseen in sample CYP4. Scattered dark-gray shale chips givethe sandstone a salt-and-pepper appearance and, alongwith other elongate clasts, create rude sedimentary layer-lng.

The earliest stages of lithic-wacke alteration cannot bedocumented because pristine lithic wacke was not found inthe cores of even the largest included fragments and the

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INcLUsIoNs

contacts of the diatreme are not exposed. The mineral as-semblage in the brown cores of the largest inclusion(sample CYP101) comprises dominant compositionally in-termedite alkali feldspar, a stilpnomelanelike phase (seebelow), and minor qtrartz. Relic clastic grains of quartz andpartially exchanged alkali feldspar preserve sedimentarybedding. Interstices between relic clasts are filled withturbid intermediate alkali feldspar, similar in composi-tion to that of the partially exchanged clasts. ThestilpnomelaneJike phase (pleochroic in amber hues) occursas xenomorphic bundles of platy or acicular grains (maxi-mum 0.5 mm) scattered throughout the matrix. This phasemay also occur, however, as delicate plumes that have nu-cleated along the margins of, and appear partly to replaceclastic quartz grains (Fig. 2A).

Boundaries between the brown cores and gray-greenrims of larger inclusions are distinct (Fig. l); this colorchange apparently reflects the presenc€ of thestilpnomelane-like phase in the core assemblage and ofacmite and alkali amphibole in the rim assemblage. Therims of the larger inclusions are mineralogically essentiallyidentical to the more homogeneous, smaller inclusions andconsist of titanian-acmite, alkali amphibole, alkali feldspar(which grades to K-rich microcline as the contact qrith thehost diatreme is approached), a natrolite-like zeolite, a fi-brous unidentified Ca-Na silicate, and rare wollastonite.The disappearance of the stilpnomelanelike phase at thecore-rim boundary is abrupt, but there is no microscopicevidence for any singular reconstitution reaction. Feldsparcompositions do not change abruptly at the core-rimboundary, and remnant clastic alkali-feldspar grains of in-termediate composition may persist well into the rims.Feldspars become progressively less sodic toward the edgeof a given inclusion, and individual grains are typicallyzoned, with rims that are enriched in K relative to theircores. Toward the interface with the matrix, feldspars insome larger zoned inclusions, and in most smaller in-clusions, are thoroughly recrystallized into mosaics of xe-noblastic to idioblastic grains of maximum microcline (asdetermined from its powder-diffraction pattern). Alkalifeldspar may display turbid cores and clear rims where itcontacts the natrolite-like zeolite (Figs. 2C and 2E). Thisdevelopment of a more nearly uniform, granoblastic tex-ture obliterates all relic sedimentary features.

Titanian acmite first appears at the core-rim boundaryof large inclusions, where chains of acnrite grains, 20 to 60pm on an edge, commonly surround relic feldspar andquartz grains (Fig. 2B). As the rim is traversed toward theinclusion/ultrahafic interfa@, acmite becomes widely dis-tributed and ever-larger idoblastic grains or clusters ofidioblastic crystals are found, although fine grains persist.The domain of alkali amphibole crystallization appears notto extend quite so far inward from the ultramafic/inclusioninterface as that of acmite, anrt alkali amphibole grains arealways larger than associateq ,.cmite crystals. The latterrelation is particularly conspicuous where there has beenmaximum crystal growth, as found in small inclusions;there, large zoned idioblastic acmite crystals (rarely 1.0 mm

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS

Table l. X-ray fluorescence analyses and CIPW norms for ultramafic host rock, lithic-wacke country rock, and metasomatized lithicwacke, Coyote Peak, California. (Bi-Shia King, analyst).

Metasomat ized I i th ic wacke

50r

Average L i th ic -wacke

Franc iscan count ry rock

graywacke* CYP4

Rim Rim t" l r l ! . lH; t " "s hostUl t r ama f i c

Core

CYP1OIA CYP1O1B CYP1O2B CYP1O3 CYPS CYPs

s i 0 2

41203

T i02

F.203

Fe0

Mn0

Mgo

Ca0

Ba0

Na20

K2o

Proq- +Hzo

Hzo-

F

67 . 5

t J . 3

0 , 5

r . 23 . 0

0 . t

2 . ?

2 , 4

3 . 6

r . 70 . 1a . a

0 . 4

70.62

13 . 11

0 .60

L 5 9

3 . 28

0 .061 A t

0 . 38

0 .04

4 . J 5

0 .82

0 . 1 3

2 . 1 4

0 .38

0 .04

61 . 02

1 6 . 1 8

0 .54

I . 84L . t a

0 . 0 6

1 . 2 8

f . i 8

0 . 7 2

4 . 3 4

7 . 7 9

0 . 10

1 . 3 8

L20

0 . 0 5

61 .98

1 5 . 9 3

0 .48

| , t l1 . 5 2

0 .06

1 . 4 8

L . O a

0 .08

? .98

1 1 . 0 8

0 . 1 1

0 . 4 9

0 . 20

0 . 1 3

61 . 07

1 5 . 3 5

0 . 4 2

2 . 6 5

0 . 4 7

0 .03

L . t )

1 . 48

0 . 0 8

3 . 4 61 1 7 7

0 . 0 2

0 .93

0 . 2 9

0 .06

J J . 6 0

9 . 4 I

3 . 14

7 .89

6 . 4 7

0 . 3 1

9 . 4 2

19 . 09

0 . 1 5

4 . 7 5

1 . 5 8

1 , 98

0 . 6 1

0 .04

0 . 19

60 .35 61 .05

15 .48 14 .74

0 .53 0 .42I 0 7 a 1 )

2 . t 7 0 . 8 4

0 .06 0 .051 0 ? ' l A ?

2 . 7 6 0 . 3 3

0 .08 0 .08? 7 ? ? q q

10.96 72.22

0 . 1 2 0 . 0 3

0 . 4 5 0 . 8 3

0 , 2 2 0 . 2 4

0 . 1 0 0 . 1 0

Total 98.70 99. 36 99 .43 99 .85 99.90 99,47 99.33 98.83*r

ac0rAb

An

Ne

Ac

N s

M9-Di

Fe-Di

Hy

0 l

Mt

I I

S a l i c

F e m i c

38 .9q ?

5 . 0? 7 0

8 . 6

2 . 4

t . 2

87 . 0

1 .2 .9

t . a

4 7 , 3

3 7 . 7

7 . 7

2 . 7

0 . 6

2 . 6

2 . 7

1 . 1

89 .0

i 0 . 9

65.920.3

4 , 4

2 . 40 , 3n o

86 .3l J - o

65.2

1 3 . 8

2 , 7

3 . 9

6 . 5

3 . i

2 . 1

0 . 9

1 . 0e l 7

1 8 . 2

7 . 7

1 9 , 8

1 8 . 3

l n q

1 1 . 1

6 . 0

29.8***

69.7r**

7 3 . 3 7 0 . 9

7 .9 11 .2

9 . 2 7 . 8

2 . 9 7 . 7

0 . 3 5 . 0

0 . 1 0 . 3

3 . 9

1 , 1 0 . 7

0 . 0 0 . 0

0 . 8 0 . 8

81 .3 83 .4

18 .7 15 .5

*Average of 21 Franciscan graywacke sandstone analyses f rom samples col lected throughout the

s e c t i o n ( B a i l e y e t a 1 . , 1 9 6 4 , T a b l e 2 , p . 3 4 ) .

* *A l so : C02 ,0 .18 ; C1 ,0 .50 i S ,0 .42 ; S r0 ,0 .24 ; and 2 r02 ,0 .09 ,

* * * I n c l u d i n g : A p , 4 . 7 l C c , 0 . 4 ; C s , 1 7 . 5 ; H ' 1 , 0 . 8 ; H m , 0 . 3 ; 1 c , 7 . 3 ; P r , 0 . 8 ; a n d 2 , 0 . 1 .

long with dark- to olive-green pleochroic cores and col-orless nonpleochroic rims) or acmite-crystal aggregates andlarger zoned xenoblastic alkali-amphibole grains (maxi-mum 1.25 mm across) occur as disseminated porphyro-blasts in a granoblastic matrix (Figs. 2D and 2E). Wherebest developed in small inclusions, alkali amphibole mostcommonly occurs as irregular glains (Fig. 2D) that can betaken at first glance for biotite; it has unusual pleochroismwith predominant light-tan to lilac cores and irregularolive-green to purple "rims." Highly serrate edges of am-

phibole grains appear to result from impingement ori smallturbid groundmass patches or acmite crystals.

Relations between acmite and alkali amphibole are dilli-cult to establish. Neither phase shows any clear indicationof instability, with acmite often displaying sharp crystalfaces and the serrate margins of amphibole grains app€ar-ing to be crisp growth features. Most acmite is not in as-sociation with amphibole, and no examples of fine acrnitecrystals including small patches of relic amphibole havebeen noted. On the other hand, many large amphibole

502 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES /N INCLUS/ONS

Fig. 2. (A) Replacement of a clastic quartz grain by the stilpnomelaneJike phase (sanple CYPT). Crystal at bortom center is alkalifeldspar. Quartz grain is 0.25 mm across. (B) Alkali feldspar and quartz surrounded by finely crystalline acmite (dark, high relief) (sampleCYPT). Feldspar grain is 0.37 mm wide. (C) Microcline, showing turbid cores and clear rims against interstitial, natroliteJike zeolite(sample CYPS). High-relief grains are acmite. It is 0.46 mm across the three large grains. (D) A typically anhedral large zonedalkali-amphibole grain in a matrix of microcline and acmite (AC) crystals (sample CYP8). Unusually large, strongly zoned acmiie crystal(right center) is 0'58 mm long. (E) A large unusually well-formed zoned alkali-amphibole crystal lsampte CYPS). NatroliteJite zeoliteoccurs both within the amphibole and between the grain and clear euhedral microcline rims. Amphibole crystal is 0.32 mm wide. Acmitegrains, AC (F) Fibrous unidentified Ca-Na silicate, intergrown with the natrolite-like zeolite on left and euhedral high relief acmitecrystals just right of center (sample CYP52). Matrix is microcline and fine acmite crystals. Largest zeolite fibers are 0.36 mm long.

(F)(E)

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS 503

grains are in marginal contact with one or more acmitecrystals or clusters ofcrystals (Fig. 2D and 2E). It is usuallydifficult to decide on textural grounds whether the acmitemore commonly was being replaced by amphibole or wasgrowing intimately upon it. This diffrculty is illustrated bythe two ragged acmite grains in the upper center and leftside of Figure 2D. Three considerations suggest that ingeneral crystallization of acmite preceded that of alkali am-phibole: (1) Acmite crystallization extended farther into thelarger inclusions, where fine crystals even coexist withquartz in zones marked by early stages of feldspar ex-change (Fig. 2B). (2) Despite their relatively large size, am-phiboles are not poikiloblastic, but typically have clear in-teriors (Figs. 2D and 2E), suggesting that they were able toassimilate or displace mineral phases, including acrnite,until their growth ceased. (3) The patchy zoning at theamphibole rims (reflecting high Ti-content) often is devel-oped in proximity to marginal acmite (e.g., Figs. 2D and28, bottom, center). Further commentary on acmite-amphibole relations is deferred to the discussion.

The natrolite-like zeolite (see below) is a common phasein the smaller inclusions, possibly less so in the rim assem-blage oflarger inclusions. It is interpreted to have been thelast phase to crystallize because it typically occurs as inter-stitial aggregates scattered within a feldspar matrix (Figs.2C and 2E).lt appears to be in stable association with bothacmite and alkali amphibole (Figs. 2C, 2E, and 2F). Thefibrous Ca-Na silicate is less common and occurs assheaves of radiating needles (maximum 0.5 mm long), oftenintergrown with the natrolite-like zeolite (Fig. 2F). Thisfibrous silicate is best developed in the smaller, more nearlyhomogeneous inclusions, but away from the in-clusion/ultramafi c interface.

Sample CYPT is atypical in that overlapping rim alter-ation zones suggest that it was somehow exposed to twoepisodes of metasomatic alteration, perhaps by becominguncoupled from its initial melt envelope. The only recog-nizable effects ofthis history are the presence ofxenoblasticwollastonite grains in parts of the rim assemblage and theoccurrence of enlarged grains of the stilpnomelane-likephase.

Between inclusions and the enveloping ultramafic host isan irregular monomineralic rind (maximum 0.3 mm thick)of light-green sodian augite (Wo.nEn.uFsruAcn; Table 2)(augite is an essential phase within the alkalic ultramafichost). Some augite grains (maximum 0.25 mm cross) mayextend from this rind as small tongues into the ultramafichost, while scattered augite grains occur within recrystal-lized lithic wacke adjacent to the augite rind and are com-monly rimmed by titanian acrnite or included within alkaliamphibole. Ilmenite and titanite are minor phases withinthe recrystallized lithic wacke adjacent to the rind; mag-netite and pyrite are rare.

Whole-rock chemistry

The analyses listed in Table I indicate the bulk chemicalchanges caused by metasomatism. Analyzed lithic-wacke

@untry rock (sample CYP4) and a representative sampleof uncontaminated ultramafic host (sample CYP5; Morganet al., 1985) are presumed to establish the chemistries of therocks prior to metasomatism. The caution in this statementis based on two reservations: (1) the fragrnents included inthe diatreme could originate from a stratigaphic intervalwhich differs mineralogically and chemically from thatpresently exposed next to it, and (2) the lithic-wackecountry rock may have been albitized by intrusion of thediatreme. We are unable to evaluate the first possibility,but consideration of data for "average" Franciscan gray-wacke (Table 1) suggests that our interpretations do notcritically depend on proof of the equivalence between theincluded and host-rock lithic wacke. Whereas a carefulsampling and analytical program could address the secondpossibility, we are unconvinced of the potential rewards ofsuch a study. Analyses of two homogeneous small in-clusions (samples CYP8 and 103), the rims of two largerinclusions (samples CYP101B and 1028), and the core zoneof one of these inclusions (sample CYPl0lA) indicate thebulk chemical changes that accompany the several mineral-ogic changes just described. Most striking is the exception-al KrO content of "fully altered" lithic wacke, whichreaches 12 wt.o/o.

Comparison of the chemical analysis of sample CYP4with an average of 21 analyses of Franciscan graywacke(Table l) suggests that this sample may be a reasonablerepresentative of the lithic wacke in the section as a wholeand thus of the sedimentary sequence penetrated by themagma during emplacement of the diatreme. The higheralkali and lower CaO contents of sample CYP4 in com-parison to those of "average" Franciscan graywacke reflectthe presence of essentially pure albite, in contrast to feld-spars elsewhere in the Franciscan assemblage, which rangein composition from albite through oligoclase or andesineand have been reported locally to be as calcic as lab-radorite or bytownite.

Mineral chemistryMineral grains were analyzed in six Coyote Peak lithic-wacke

samples, but most of the data presented here were obtained from

samples CYP4, CYP8, CYP52, and CYP101. Mafic silicates inpolished thin sections were analyzed with a three-channel ARLEMX-SM electron microprobe, using an accelerating potential of15 kV and a sample current of 0.O2 pA on benitoite. Elementswerc analyzed in the sequence Mg * K + Fe, Al + Si + Ti,F + Cl + Ca, and Na + Mn. All analyzed grains were photo-

graphed to allow relocation to be as precise as possible. Eachreported analysis represents 6 to 12 points per grain or distinctportion of a zoned grain; each point was occupied for a countinterval (about 10 seconds) fixed by beam-current termination.The selected amphibole, biotite, and pyroxene described by Cza-manske and Wones (1973) were used as standards for most ele-ments in the appropriate mineral species. Supplementary stan-dards were MnrO. for Mn, TiO, for Ti, a synthetic fluor-phlogopite for F (9.01 wt.Vo), a natural crocidolite for Na (6.24

wt.'h), and a natural sodalite for Cl (6.82 wt.%). Natural albite,bytownite, and orthoclase were used as standards for the feldsparanalyses. Count data were stored, and analyses computed online,using the data-reduction scheme FRAME (Yakowitz et al., 1973).

504 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN /NCLUSIONS

Table 2. Electron-microprobe analyses and structural formulae for representative acrnite and augite grains

ACI.4ITE AUGITE

L i t h i c - w a c k e i n c l u s i o n s

C Y P 8 . B

Z e o l i t i c

c l o t s

( range o f 7 )

L i t h i c - w a c k e i n c l u s i o n s U l t r a m a f i c

C o r e R i n Core R im

CYP52, I I C Y P 5 2 , I V

C o r e R i m

C Y P 8 . I ]

R i m

CYPlO1 , CYPB,V I I D

nost

( range o f 28)

s i 0 2

T i 0 .

A t r 0 1

r"6r"Fer0 lC

r"6c-Mn0

!190

Ca0

Na20

Kzo

Tot a l *

s iA I

T e t .

r,r( 1 )

' 'oct

N a

M ( 2 )

Fe/ ( Fe+t4g )

5 2 . 9 5 2 . 5

8 . 3 8 I . 2 4

0 . 3 2 0 . 3 9

27.8 26 ,9

1 5 . 6 6 ? 9 . 9 3

7 . 6 8

0 . 0 5 0 . 0 5

f . i 8 1 . 2 4

1 . 1 1 0 . 9 3

1 2 . 6 1 3 . 0

5 3 . 0 5 4 . 4

I 0 . 0 6 . 9 1

0 . 3 1 0 . 2 8

1 9 . 9 2 2 . 1

1 3 . ? 2 1 9 . 4 8

8 , 0 3 4 . 6 0

0 . 0 5 0 . 0 9

I . 9 7 2 . 0 9

0 . 9 8 0 . 6 2

I 2 . 8 1 3 . 1

5 3 . 2 5 3 . 9

5 . 8 3 0 . 1 2

0 . 3 4 0 . 7 1

22.8 28 ,3

2 2 . 7 9 3 1 . 4 5

2 . 2 4

0 . 0 1 0 . 0 1

2 . 3 4 1 . 6 6

1 . 0 3 0 . 9 3

1 3 . 2 1 3 , 2

0 . 0 2

53.2 52 .9

7 . 2 1 0 . 7 ?

0 . 1 8 0 . 4 7

2 2 . 8 2 9 . 3

1 9 . 1 6 3 1 , 8 1

5 . 5 4 0 , 6 5

0 . 0 5 0 . 0 2

1 . 7 4 1 . 0 1

0 . 2 1 0 . 1 l

1 3 . 1 1 3 . 2

5 1 . 6 - 5 3 . 2

1 . 6 4 - 4 , 4 1

o . 2 2 - 1 . 2 7

24.O - 3 I .7

20.9 - 34.9

0 . 2 7 - 5 . 2 3

0 . 1 6 - o . 2 2

0 . 0 0 - 2 . 4 0

0 . 1 4 - 3 . 8 6

1 1 . 3 - 1 3 . 4

52.8 53 .1

0 . 6 8 0 . 8 5

o . 4 2 0 . 6 8

1 1 . 7 1 0 . 7

1 . 9 2 2 . 3 ? .

9 . 9 6 8 . 6 I

o . 4 2 0 . 3 1

1 ? . 7 t ? . O

1 9 , 3 2 0 . 6

I . 2 8 1 . 6 5

5 1 . 3 - 5 5 . 9

0 . 1 9 - 1 . 4 1

0 . 1 8 - I . 5 7

3 . 8 0 - 1 0 . 1 1

0 , I 7 - 4 . 8 2

i . 7 4 - 8 . 3 6

0 . 0 8 - 0 . 4 6

1 1 . 9 - 1 5 . 9

2l 2 - 25,4

0 . 5 1 - 2 . 1 5

100.44 99 .32 100.36 10 i .57

1.994 2 .O08 1 .991 2 .014

0.006 0 .009

2 . 0 0 0 2 , 0 0 8

0 . 0 0 9 0 . 0 1 8

o.444 0 .862

0 . 2 4 2

0 . 1 0 0 0 . 0 7 1

0.002 0 .002

0 . 2 3 7 0 . 0 3 6

2.000 2 .014

0 . 0 0 5 0 . 0 1 2

0 , 3 7 4 0 . 5 4 3

o . 2 5 2 0 . 1 4 3

0 . 1 1 0 0 . r 1 5

0.002 0 ,003

0 . 2 8 3 0 . I 9 2

0 . 0 3 1

0.642 0 .882

0 . 0 7 0

0 . 1 3 0 0 . 0 9 2

0 . 1 5 4 0 . 0 0 3

0 . 0 0 8 0 . 0 2 1

0. s43 0 .906

0 . 1 7 4 0 . 0 2 0

0 . 0 9 8 0 , 0 5 7

0.002 0 .001

o.204 0 .020

99.48 99 .82

1 .991 1 .988

0 . 0 0 9 0 . 0 1 2

2.000 2 .000

0 , 0 1 0 0 . 0 1 8

0.05s 0 .065

0 . 3 1 4 0 . 2 7 0

0 , 7 1 5 0 . 6 6 7

0 . 0 r 3 0 . 0 1 0

0 , 0 1 9 0 . o 2 4

1 . 1 2 6 1 . 0 5 4

0 . 1 2 6 0 . 0 5 4

0 . 7 8 0 0 . 8 2 6

0 , 0 9 4 0 . 1 2 0

100.98 102.00 100.39 100.79

1.985 2 .008 ? .O02 2 .001

0 . 0 1 5

2.000 2.008 2.002 2.001

A I_ ? +l 'e -

F e '

Mg

Fln

t . 0 3 3 0 . 9 8 9 1 . 0 2 6 1 . 0 0 8 1 . 0 0 6 1 . 0 0 8 1 . 0 2 9 1 . 0 2 5

0.033 0 .026 0 .008

0 . 0 4 5 0 . 0 4 0 0 . 0 3 9 0 . 0 2 5

0 . 9 2 2 0 . 9 6 ? 0 . 9 3 5 0 , 9 4 0

0 , 0 0 6 0 . 0 0 8 0 . 0 2 9 0 . 0 2 5

0 . 0 4 1 0 . 0 3 7 0 . 0 0 8 0 . 0 0 d

0 . 9 5 4 0 . 9 5 6 0 . 9 5 9 0 . 9 6 8

1 . 0 0 0 1 . 0 0 2 1 . 0 0 0 0 . 9 7 3

0 . 8 7 3 0 . 9 2 4 0 . 8 5 0 0 . 8 5 6

1 . 0 0 1 1 . 0 0 1 0 . 9 9 6 0 . 9 9 8

0.845 0 .906 0 ,880 0 .942

0 . 3 7 4 0 . 5 5 2

0 . 0 0 5 0 . 0 1 2

0 . 2 2 0 0 , 2 3 3

0 . 3 3 6 0 . 1 5 6

0 . 0 0 5

0 . 0 0 2 0 . 0 0 3

0 . 0 1 6 0 . 0 1 I

0 .042 0 .033

o.642 0 .898

0 . 0 3 1

0 . 2 6 0 0 . 0 0 6

0 . 0 5 2

0 . 0 0 7

0 . 0 1 7 0 , 0 1 9

0.046

o . 0 2 2

0.544 0 .904

0.008 0 .021

0 . 1 9 6 0 . 0 4 0

o.2L3

0.002 0 .001

0 , 0 0 3 0 . 0 0 2

0.o29

0.034 0 ,003

i . 0 0 0 1 . 0 0 0

0 . 3 4 0 0 , 3 3 4

0 . 0 5 5 0 . 0 6 5

0 , 0 1 0 0 . 0 1 8

0 . 0 2 9 0 . 0 3 7

0.005 0 .006

0 . 0 1 3 0 . 0 1 0

0 , 3 8 2 0 . 4 0 5

0 . 3 4 9 0 . 3 2 4

0 . 1 5 7 0 . 1 3 5

l laFe3*s i rou 0 .444 0 .871

N a A l 5 i 2 0 6 0 . 0 0 9 0 . 0 1 8

N a M g . 5 T i . 5 S i 2 0 6 0 . 2 0 0 0 . 0 7 3

N a F e . 5 T i . 5 5 ' i 2 0 b 0 . 2 6 9

CaTiA l206 0 .003

CaMnSi206 0 .002 0 .002

C a 2 s i 2 0 s 0 . 0 2 0 0 . 0 1 9

r , !92s i206 0 .017

F e 2 S i 2 0 6 0 , 0 5 3

* Inc ludes FerOrC and Fe0C.

Fe0 l cor responds to mic roprobe de ter r ina t ion o f to ta l Fe .Fe203" and Fe0" a re va lues genera ted by the conputer p rogram o f Pap ike e t a l . (1974) .

Pyroxenes

The 10 selected pyroxene analyses listed in Table 2 rep-resent a total of 22 acmite and 7 augite analyses from foursamples. High sums for two of the acmite analyses appar-ently relate to difliculty in reoccupying portions of thesestrongly, and nonuniformly, zoned crystals. Distribution of

Fe3+/Fe2+ and structural formulae were calculated withthe computer program of Papike et al. (1974). Normativepyroxene components were calculated according to a pro-cedure proposed by Malcolm Ross, U.S. Geological Survey(written communiation, 1982). Relative to the cores, rims ofacmite grains are enriched in Fe and depleted in Ti; therims also generally contain slightly more Al and less Ca

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS 505

and Mg. Fe/(Fe + Mg) ranges from 0.830 to 0.978, andNa/(Na + Ca) from 0.953 to 0.995; both ratios are alwaysgreater in the rims of grains. In agreement with Ferguson(1977), our data do not support Flower's (1974) proposalthat Al, Fe, and Ti substitute significantly for tetrahedral Siin titanian acmite.

Total measured ranges (in weight p€rc€nt) in the compo-sition of acmite within lithic-wacke inclusions are: SiOr,5 1.8-54.5 ; TiO2, 0. 12-10.2; AlrO r, 0. 18-1.32 ; FeOr, 18.5-29.3; MnO, 0.0GO.20; MgO, 0.3G2.53; CaO, 0.11-1.36;and NarO, 12.G13.3. As listed in Table 2, this compo-sitional range resembles that shown by acmite crystals thatare associated with natrolite, sodalite, and pectolite inzeolite-rich clots (several crn across) within the comtamina-ted ultramafic host. A notable exception is TiOr: Ti con-tents in the cores of some acmite grains within lithic-wackeinclusions may establish a new maximum (0.297 cationl6oxygen ions; analysis not reported in Table 2) for pyrox-enes (Robinson, 19E0). Acmite in both environments mayclosely approach the ideal composition, NaFe3*SirOu;SiO2-52.02, FerOr-34.56, Na2O-l 3.42 (weight percent).

Also listed in Table 2 are compositions of the augite thattypically forms a rind at the contact of inclusions with theultramafic host but may also occur as discrete grains (max-imum 8 mm) within inclusions. For comparison, we showthe compositional range of the augite that is an essentialconstituent of the ultramafic host. There are significant dif-ferences between the two augites in the contents of Ca, Fe,and Mg. While it was expected that the augite in the ultra-mafic host would contain more Ca and Mg, it is surprisingto find that it contains less Fe.

Figure 3 illustrates the strong compositional zoning be-tween the Ti-rich dark-green cores and the Ti-poor Fe-enriched colorless rims: Ti content is plotted against totalFe content and Fe2+/(Fe2* + Fe3*) for all acceptableanalyses. Figure 3A shows a good inverse correlation,averaging 1:1, between Ti and total Fe contents; Figure38 shows that this exchange is well accounted for by vari-ations in Fe2*l(Fe2* +Fe3+; within the crystals. Ex-change of Tia+ + Fe2+ within the cores for Fe3+ withinthe rims may be regarded as substitution within the solid-solution series NaFe3 * SirOu-Na(Fe2 *,Mglo.rTif

.:Si206(Ferguson, 19771' Deer et al., 1978, v. 24, p.486, 493). Thereference lines in Figure 3 are for a specific case in whichone-third of a total of 0.9 Fe cations was initially Fe2+,coupled with Tia*. Poor correlation of Ti content withFe/(Fe + Mg), indicates that Mg is not fundamentally in-volved in the coupling and substitution mechanism (notethe moderate and relatively uniform MgO content of theseacmites).

Relations between Na and Ti in pyroxenes are complexas is evident from the data of Flower (1974, Fig. 6) andRonsbo et al. (1977, Fig. 2). At Coyote Peak, marked deple-tion of Ti from core to rim of acrnite grains is typicallyaccompanied by only slight enrichment in Na (Table 2);this is in agreement with Ferguson (1977, p. 250), whonoted that minor Ca may substitute for Na with increasingTi content, although the zoning pattern he referred to is the

0 6 5 0 7 5 0 S 5 0 9 5Fe, colions/6 O

o m L0 5 5

o 0 2 0I

oo

o t o

o

oo

\

\

o20

0 t 0

0 2 0 ooo0 t 0F6z+ /(Fe2++Fe3.), olomic

Fig. 3. (A) Relation between Ti and total Fe contents in zoncdpyroxenes. (B) Relation between Ti content and atomicFe2'/(Fe2* + Fe3+; in zoned pyroxenes.

reverse of that observed in Coyote Peak acrnite. Continu-ing loss of Ab component from adjacent feldspars mayexplain the continued availability of Na at the time of rimcrystallization.

Although the number of analyses is not great, we notethat pyroxene cores in sample CYP8, the small homoge-neous inclusion, are lower in Tia+ and Fe2+ contents thanthe pyroxene cores in sample CYP52, a medium-size zonedinclusion. As discussed below, amphibole chemistry showseven sharper, size-dependent variation.

Because titanian acmites have recently received muchattention (e.g., Ferguson, 1977; Ronsbo et al., 1977; Niel-sen, 1979), it seems appropriate to reflect briefly on no-menclature and calculation schemes for these pyroxenes. Inaddition to the calculations listed in Table 2, we have recal-culated the 7 analyses in Table I of Ferguson (1977), the 7analyses in Table 2 of Ronsbo et al. (1977), and the 10analyses in Table 4 of Nielsen (1979). We have used theend-member components and the order of calculation(Table 2) proposed by Malcolm Ross (written communi-cation, 1982). Of 34 analyses, only 2l calculate out to giveresiduals (i.e., cations which cannot be allocated to an endmember) of less than 0.020 cation, and these give anaverage residual of 0.(XX cation; larger residuals for otheranalyses suggest that those analyses are inferior. For exam-ple, analyses of three rims of gxains from Coyote Peakcalculate out with residuals (in cations) as follows:CYP52,II,R-Na, 0.010 and Si, 0.027; CYP52,IV,R-Si,

o

506 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIB,LES IN INcLUsloNs

0.042; and CYP8,II,R-Na, 0.037 and Si, 0.044. Becausethe cores of these grains give an average residual of 0.001cation, we suspect that crystal heterogeneity and uncer-tainty in beam relocation caused the residuals for the rims.For analysis 7 of Ronsbo et al. (1977) and analysis 6 ofNielsen (1979), residuals of 0.043 and 0.023 cation werereduced to 0.002 and 0.001 cation, respectively, by insertionof the component NaTiAlSiOu in lieu of CaTiAlrO.; foranalyses 7 and 8 of Nielsen, large residuals were cut in halfby use of this component. However, unacceptable residualsof 0.048 to 0.251 cation remain for the latter two analyses,for analysis 4 of Ronsbo et al., for analyses 1, 4, 5, and 6 ofFerguson, for analyses 2 and,3 in Table 5 of Ronsbo et al.,and for analysis 19 in Table 53 of Deer et al. (1978, p. a9\.Therefore, we believe that many reported analyses of tita-nian acrnite are in error and that care must be exercised intheir acceptance. The component NaTiAlSiO. seems to beof questionable significance in these pyroxenes.

Figure 4 is a plot of normative components for the py-roxene compositions listed in Table 2 and the apparentlysuperior analyses of Ferguson (1977), Ronsbo et aL. (1977),and Nielsen (1979) on a modified pyroxene ternary dia-gram in which the end member Na(Fe2+,Mg)o.5Tio.5Si2O6is placed at the apex. Analysis I ofFerguson is also plotted,despite identical large residuals of 0.024 cation for Fe2*and Ti, because it is the most Ti-rich acmite heretoforereported. Coyote Peak acrnite is apparently unique in fol-lowing an extended trend of low and nearly uniform molefraction of "Quad"-plus-"Other" components.

It is significant that for the first time a composition hasbeen determined for a pyroxene which contains more than50 mole% of the end member Na(Fe2*,Mg)o.5Tio.ssi2o6.Although we do not undertake here the task of naming thisend member, we note the following five points: (1) It seemsmost unfortunate that the term "neptunite" waschampioned for this end member by Ferguson and takenup by Deer et al. (1978, p. 486, 493). Neptunite is not apyroxene, and we strongly discourage use of this name. (2)In our opinion, the substitution of Ti4+ is of fundamentalsignificance in setting this end member apart from acmite.Thus, we do not favor establishing two end members,namely, NaFefr.+rTio.rSirOu and NaMgo.rTio.rsi2o6. (3)Moreover, because the mole fractions of these two compo-nents, commonly referred to as Fe2*-NAT and Mg-NAT(e.g., Cameron and Papike, l98l), generally depend signifi-cantly on the order ofcalculation and thus can be mislead-ing, we recommend that discussion in these terms be aban-doned. (4) Failing that, and in consideration of our datasummarized in Figure 3, we suggest that there is goodcause to reverse the common order of calculation for thesepyroxene components containing M(1FM(2) cations offour valencies (Na+, Fe2+, Fe3+ and Tia*), such that thecomponent NaFee.*5Tio.5sirOu is calculated before thecomponent NaMgo.rTio.rsiro6. (5) Following commonusage, these pyroxenes might best be called simply "tita-nian acmite" until the end member is established as a newmineral species.

Fig. 4. An improvised pyroxene ternary diagram in which thecomponents Na(Fe2 +,Mg)o.sTio.5si2o6 and NaFe3 +SirOu havebeen isolated and all other components (see Table 2) have beencombined with the "Quad" components at a third corner. Sampledesignations are the same as in Table 2, and, for the other refer-ences are from respective tables cited in text.

Amphiboles

A total of 47 amphibole analyses were obtained, repre-senting 29 grains in samples CYP7, CYP8, CYP52,CYP101, and CYP178. In most cases, core and rim compo-sitions were determined for these conspicuously, but non-uniformly, zoned grains (see Figs 2D and 2E). Table 3 listsdata for representative analyses, and data for all analysesfrom samples CYP8, CYP52, and CYP101 are included ina series of plots (Figs. 5, 6, and 7) to depict the substitutionrelations found in these alkali amphiboles. Table 3 alsolists structural formulae, and "midpoint" values of Fe2O3and FeO calculated with the amphibole calculation pro-gram of Papike et al. (1974). Because of the heterogeneity,small size, and intergrowth of these amphibole grains, wehave not pursued their optical and X-ray properties.

Typically, the amphibole cores are ferroan richterite, andthe rims titanian magnesio-arfvedsonite (Leake, 1978;Hawthorne, 1981). Low Al content leads to a designationas arfvedsonite rather than eckermannite. Considering theiruniformly low AlrO, contents, and CaO contents that mayfall below I weight percent, the arfvedsonitic rims appearto be among the nearest to ideal composition yet reported,with one caveat-they contain Tia+ rather than Fe3+. (Infact, we use the term "arfvedsonite" with misgiving andbelieve that eventually the designation of a titanianalkali-amphibole end member, NaNar(Fe,Mg)4.5Tio.5Si8O22(OH)2, may be justified.) Their zoning thus provides afascinating contrast to the pyroxene zoning. Most analysesrepresent compositions that should properly be prefixed"ferroan," "fluor-," and (or) "titanian"; for simplicity, wesometimes refer to these amphiboles as richterites.

Most of our analyses contain insufhcient Si and Al to fill

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES /N /NCLUSIONS

Table 3. Electron-microprobe analyses and structural formulae for representative amphibole grains in lithic-wacke inclusions andzeolitic clots

C Y P 8 , I ] cYP8 ,V cYP52 ,V cYP52 , IX cYP10 l , I l cYP 101 , V

507

Core R im Core R i m Core R im C o r e R i m C o r e R i m Core R im

Z e o l i t i cc l o t s

( r a n g e o f 5 )

s i 0 2T i02or19,F e O '

Fer01C

r"6c-! ln0

M9o

Ca0

Na20

K2o

F*

5 5 . 0

2 . 4 6

0 . 2 9

8 . 4 7

8 . 4 7

0 . 0 6

r / . o

5 . 6 4

b . r /

I . 6 3

2 . 9 5

0 . 3 77 . 89

7 .89

0 . t 4

1 9 . 05 . 1 9

6 . 0 8

1 . 4 33 . 6 6

5 3 . 06 . ? l

0 . 2 5r 7 . 9

1 7 . 9

0 . 2 59 .444 . 5 2

6 . 5 ?

1 . 5 51 . 2 4

5 6 . 0

2 . 0 2

0 . 4 0

9 . 3 4

9 . 3 4

0 . 2 0

1 7 . 8

5 . 9 9

6 . r 71 . J O

3 . 4 1

52.9

6 . 39

0 . 1 92 3 . 8

23.8

0 . 2 34 . 7 3

? . 6 6

5 . 9 3

1 . 5 21 . 0 1

5 5 . 8 5 4 . 62 . t 9 2 . 9 5

0 . 3 8 0 , 2 47 . 7 6 1 3 . 8

0 . 8 77 .76 12 .99

0 . 0 4 0 . 1 21 7 . 8 1 3 . 94 . 6 1 1 . 6 0

6 . 5 2 8 , 1 9

7 . 6 7 r . 7 ?? . 6 0 1 . 3 1

5 5 . 5 5 4 . 3

2 . 2 1 3 . 5 9

0 . 2 9 0 . 1 99 .93 20 .1

0 . 5 6 1 . 3 9

9 .43 18 .810 . 1 1 0 . 1 3

17 . I 8 . 933 . 1 9 0 . 4 9

7 .22 8 .82

1 . 5 8 L . 7 l1 .94 0 .88

5 4 . 1 5 5 . 5 5 4 . 5

3 . 4 9 3 . 1 3 4 . 8 7

0 . 1 1 0 . 3 0 0 . 2 4L r . z? 9 .58 r7 . 5

11 .22 9 .58 r7 . 50 .06 0 .09 0 .09

1 5 . 3 1 6 . 7 1 0 . 34 . 6 7 5 . 1 5 2 . 2 6

6 . 6 1 6 . 4 2 7 . 7 9

1 .60 1 .60 1 .582 . 0 1 ? . 2 0 1 . 1 3

T o t a l * * 9 8 . 2 8 9 7 . 9 4 9 8 . 3 1 9 8 . 9 7 99 .93

7.928

0 .0340 .038

8.000 8.000 8.000 8.047 8.000 8.000 8.000 8.000 8.000 8.000 8.000 8.000

s iAIT i

A 1Fe"

,+Fe '

14g

Mn

T i

M ( 1 ) - M ( 3 )

^oc t

C a

N A

! r ( 4 )

N a

N

0 .043 0 .0160 .095

0 . 9 2 7 1 . 5 8 7

3 .78? 3 .02s0 .005 0 .0150 .235 0 .324

0 .010 0 .033

0 .060 0 .155

1 . 1 3 2 2 . 3 3 03 .650 1 .971

0 . 0 1 3 0 . 0 1 60 .239 0 .400

1 .014 1 .364

3 . 763 3 . 3100 .007 0 .0070. 186 0 .26L

1 . 138 2 .L39

3 .536 2 .226

0 . 0 1 i 0 . 0 1 10 .264 0 .521

0 .939 2 . t 914 .0?3 2 .0590 . 0 1 7 0 . 0 3 10 .070 0 .481

1.099 ?.9893 . 7 3 5 1 . 0 5 6

o .024 0 .0290. 1s4 0 .682

7 .979 7 . 975 7 .961 8 .0470 .021 0 .025 0 .039

99.03 98.32 99.74 99.78

7 .872 7 .860 7 .879 7 .9460 .049 0 .019 0 .050 0 .04 I0 .079 0 ,121 0 .071 0 .013

99. 36 100. 36 101 . 2s

7 .843 7 .755 7 .874

0 .062 0 .043 0 .066

0 .095 0 .?02 0 .060

4.992 5.062 5. 104 4.905 4.970 4.942 4.949 4.897 5 .049 4 .762

0 .0490 .944 0 .7091 .007 1 .29L

5 .012 4 .156

0 .0120 .903 0 .427

1 . 0 8 5 1 . 5 7 3

0 .062 0 .0140 .706 0 .251 0 .491 0 .078t . 294 1 .687 1 .405 1 .922

0 .865 0 .727 0 .784 0 .3531 .135 1 .273 1 .2 t6 ! . 647

2.000 2.000 2.000 2.000

0 .51? 0 .632 0 .606 0 .611

0 .304 0 .321 0 .289 0 .323

2.000 2.000 2.000 2.000

0 . 5 7 6 0 . 5 9 1 0 . 5 5 3 0 . 5 5 s

0 .298 Q .297 0 .290 0 .294

2 .000 2 .000

0 .670 0 .559

0.260 0.289

2.000 2.000

0 .598 0 .441

0 .244 0 .291

A 0 . 8 1 6 0 . 9 5 3

Na/ ( Na+K ) 0 . 856 0 .878

F e / ( F e + t ' 1 9 1 0 . 1 9 7 0 . 3 5 7

RIC- I1AX 0 .706 0 .251

ARF-MAX 0 .470 0 .648

0 .895 0 .934

0 .874 0 .887

0 .246 0 .558

0 .491 0 .078

0 .478 0 .795

0 .874 0 .888

0 .852 0 .863

0 . 2 1 2 0 . 2 9 2

0 . 8 6 5 0 . 7 2 7

o .372 0 .522

0 .843 0 .849

0 .859 0 .882

0 .243 0 .490

0 .784 0 .353

0 .528 0 .8?4

0 .930 0 .848

0 .866 0 .865

0 . 1 8 9 0 . 5 1 6

0 .930 0 .709

0 .140 0 .646

0 .84? 0 .732

0 .873 0 .874

0 .227 0 .739

0.842 0.427

0 .308 0 .732

* N o C l i s p r e s e n t .

* *Based on ca lcu la ted FeZ03 dnd Feo va lues and cor rec ted fo r 0 equ iva len t to F .

the tetrahedral sites (see Table 3). On the basis of extensiveexperience with our analytical procedure, we consider thisdeficiency to be real and have filled the tetrahedral siteswith Ti (Papike et al., 1969; Charles, 1977). The deficiencyin tetrahedral Si in these amphiboles is compatible withErnst's (1962) discovery that arfvedsonitic amphiboles are

silica deficient if formed at oxygen fugacities lower thanthose defined by the hematitrmagnetite buffer. Ernst sug-gested that Fe3* may replace tetrahedral Si in arfvedsoni-tic amphiboles formed at low oxidation states. Charge andsimilar effective radius (Whittaker and Muntus, 1970) makeTia* an equally or more acceptable replacement for Sia+

508 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIB,LE| IN TffcLUsIoNs

in these amphiboles, in which Fe3 + is an insignificant com-ponent.

Comparison with the compositions of amphiboles foundwithin zeolitic clots in the contaminated ultramafic hostshows that amphiboles in the lithic-wacke inclusions havehigher Ti and lower Al contents, a result opposite to whatmight have been expected, based on the whole-rock com-positions of samples CYP4 and CYp5. The amphiboleswithin the zeolitic clots also are notably richer in K andMn and do not range so widely in Mg content; none con-tain so much Mg as do the cores of grains within theinclusions.

Charles (1975, 1977) studied the phase equilibria of end-member and intermediate compositions in the systemNarCaMgrSisO22(OH)2-NarCaFerSirO rr(OH), undervaryingfo,-T conditions and concluded that the richterite-ferro-richterite series represents a complete solid solutionbetween end members. This conclusion is consistent with aplot of Fevr versus Mgvr in the amphiboles from CoyotePeak (Fig. 5), in which there is a nearly continuous substi-tution of Fe for Mg over the rather extensive range ofanalyznd, complex amphiboles.

Papike et al. (1969) indicated that two alkali ions couplewith a single Ca2* ion in the richterite lattice, such that anequal number of Na and Ca cations occupy the M(4) siteand Na occupies the A site. A plot of NaM(a) versus Cacontent (Fig. 6) supports the idea of Na substitution forCa: richterite cores are relatively enriched in Ca and ap-proach Na: Ca : I : I in the M(4) site, whereas correspond-ing arfvedsonitic rims are enriched in Na relative to Ca andapproach Na:Ca:2:0 in the M(4) site. We cannot en-tirely explain the deviations from ideal Na-for-Ca substitu-tion reflected in Figure 6. In most cases, however, the de-

0.8 1 .0 1 .4 t .8 2 .2 2 .6 3 .0Fevt, ions/23 O

Fig. 5. Relation between Mg and Fe occupancy of octahedralsites in zoned alkali amphiboles.

\. \ \o \

t . 0

ocr)c1oco

= '0 .4o

(J

o

o(v,

C-,1

oc.9

o):

1.0 t .2 t .4 1 .6 L I 2 .ONoMa, ions/23 O

Fig. 6. Relation between Ca and Na occupancy of M(4) site inzoned alkali amphiboles.

viant points correlate with calculated octahedral-site oc-cupancies of over 5 cations and thus may be an artifact ofthe calculation scheme.

Because the classification of alkali amphiboles does notyet depend on TiO, content, the extreme compositionsmay be regarded as those listed in Table 3 for the core ofgrain CYP101,II and the rim of grain CYP8,V. Corre-sponding formulae, showing the shift in Na and Ca occu-pancy of the M(4) site, are:

(Ko.r.Nao.urfiar.orCao.nnMno.orMBo.orXMgr.""Fefr .!,nTio.or)

(si7.s4Alo.o6Tio.1o)o22(oHo.35F1 6s)

and

(Ko..rNao.. r[Na,.rrCao.or(Mgr.e?F€;.:3Feel6Tio.noAlo.orMno or)

siE.oso22(oHl r"Fo.or).

Clearly reflected in these two formulae is the relative con-stancy of A-site occupancy, in contrast to the substantialdifferences in constitution of the M(4) and M(1fM(3) sites.The work of Huebner and Papike (1970) suggests completesolid solution between K and Na in the A site of the rich-terite lattice. Consideration of Table 3 shows that the Kcontent of the A site varies little (0.22-{.32 cation for thefull data set) and that A-site Na content may increase ordecrease from core to rim for individual grains.

Ti is a significant element in amphibole substitutionmechanisms, as shown by many recent microprobe analy-ses of arfvedsonitic amphiboles occurring with titanianacmite in peralkaline and undersaturated rocks (e.g., Scott,1976; Ronsbo et al., 1977; Grapes et al, 1979; Nielsen,1979). However, none of these analyses approaches the

1 . 0

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS 509

Coyote Peak rim compositions with respect to Tia* con-tent, which may reach 0.7 cation per formula unit; Ti ap-pears partly to fill the tetrahedral site and totally supplantsFe3+ in the M(1lM(3) sites. Figures 7A and 78 show goodcorrelations between TivI content (cationsl23 oxygen ions)and both Fe/(Fe + Mg) and Na/(Na * Ca). Whereas thedata of Table 3 can be taken to suggest that total Ti, Fe,and Na contents are related, poorer correlations than thoseof Figures 7A and 7B result when TiIv*vr content is plottedversus Fel(Fe + Mg) and Na/(Na + Ca), indicating thatTivr is the more significant variable (as well as supportingassignment of Tia* to the tetrahedral site). We note thatbecause Fe3* is not a significant component in these am-phiboles, the charge-balancing mechanism is apparentlyTi4* + Na* for Mg'* + Ca2* and should not be intrinsi-cally sensitive to oxygen fugacity. The Ti-rich Coyote Peakamphiboles are of special interest in view of considerablediscussion over the mechanisms of incorporation of Fe3+and Ti"* into the amphibole structure (e.g., Whittaker,1949 and 1960; Ghose, 1966; Charles, 1974; Kitamura etal., 1975). Table 3 reveals poor correlation among: (1) oc-cupancy of Tia* in the M(1fM(3) sites, (2)Na in the M(4)site, and (3) A-site occupancy. Our data do not support theinverse relations based on charge balancing that have beenproposed for accommodation of Fe3 + and Ti4 + .

Calculation of amphibole norms in the manner present-ed for the pyroxenes requires too many assumptions re-garding end members and priorities of calculation. How-ever, for comparative purposes only, we have calculatedinilepenilently the maximum mole fraction of each of thetwo end members (Na,K)CaNa(Fe,Mg)rSirOrr(OH), and(Na,K)Nar(Fe,Mg)n.rTio.rSi8O22(OH)2 which can bederived from the structural formulae listed in Table 3:these values are listed at the foot of Table 3 as nrc-uex andARF-MAx. A notable result of attempts to calculate am-phibole end members was the realization that the rims ofgrains CYP52,IX, CYP101,II, and CYP101,V contain sub-stantially more Ti than can be accommodated by the pro-posed titanian-arfvedsonite end member. Moreover, Table3 shows that the calculated occupancy of M(1)-M(3) sites issignificantly less than 5 for the Ti-enriched compositions.The existence of such electrochemically balanced, butcation-deficient components as (Na,K)NarFe3TiSi8O22(OH), and []CaNa2Fe.TiSi8O22(OH)2 must also be con-sidered.

Strikingly apparent in Figure 7B are the distinctchemical-evolution paths that characterize amphibolesfrom individual inclusions and that apparently are a func-tion of inclusion size. These trends are even better definedin Figure 7C, in which Fe/(Fe + Mg) is plotted againstNa/(Na + Ca). Both Figures 78 and 7C show that am-phibole cores within the three inclusions are relatively simi-lar in composition but evolve to strikingly difierent rimcompositions.

Fluorine, an element that has a large afect on amphibolestability, varies significantly in concentration from core torim within the amphiboles and with respect to Fe/(Fe + Mg) (Fig. 7D). The 03 site in amphiboles is coordi-

nated by one M(3) and two M(1) cations. Cameron andGibbs (1973) showed that entry of F into this site reducesthe size of the M(1) and M(3) sites and of the octahedrallayer as a whole. Rosenberg and Foit (1977) showed thatFe2* is not favored in F-coordinated sites due to smallercrystal field-stabilization energy. Thus, the more F- andMg-rich amphibole cores can represent a more thermallystable phase characterized by relatively strong Mg-Fbonds and restriction of Fe2* to the M(2) site. The am-phiboles evolve toward compositions of lower thermal sta-bility containing less F and more Fe (Fig. 7D). A plot of Fcontent versus Na/(Na + Ca) (Fig. 7E) supports data pre-sented elsewhere which show that richterite cores, enrichedin F + Ca, also evolve into rims enriched in Na and low inF. Figures 7D and 7E reveal that F content in the growingamphiboles also appears to have been a function of in-clusion size. As for the pyroxenes, amphibole cores in thesmallest inclusion (sample CYP8) are more "evolved" thanthose in the larger inclusions; they contain less F (Figs.7Dand 7E) and have a higher Na/(Na + Ca) (Figs. 7B,7C,and 7E).

The zeolite-rich clots within the contaminated ultramafichost are thought to have crystallized from a late-stagefluid-rich phase. The similar compositions of amphibolesand pyroxenes within these clots and the lithic-wacke in-clusions result not from any relation between the two fea-tures, but from a common environment imposed by theultramafic host.

Feldspars

The analyses listed in Table 4 represent 34 analyses offeldspars within lithic-wacke inclusions and a sample of theintruded Franciscan assemblage (CYP4) taken a few metersfrom the intrusive diatreme contact. Recall that the ultra-mafic host is nepheline bearing, and contains no feldspar.

Sample CYP4 contains relic plagioclase grains of near-ideal albitic composition (Ab"n..Oro.rAno.r) that are quitehomogeneous, considering that the analyses include coreand rim compositions for six grains. This plagioclase isunusually sodic for Franciscan graywacke (Bailey et al.,1964\, and it is possible that intrusion of the diatreme albi-tized this plagioclase without affecting rock textures.

As discussed earlier, it seems diffrcult to prove that thelithic-wacke inclusions were identical to the lithic-wackewallrock at the present level of exposure. We perhaps makethat assumption most explicitly in the following discussion,in which we interpret our analyses to establish a progres-sive, but commonly aborted, exchange reaction whereby Kwas exchanged for Na within relic feldspar grains. Feldspargrains in the cores or at the core-rim interface of largerinclusions have a different range of compositions fromthose found in more uniformly altered small inclusions ornear the margins of larger inclusions. Plagioclase in thecore assemblage typically is strongly zoned, with K['{ahigher in the rims of grains (Table 4). The examples chosenfor Table 4 represent extremes of KrO content; in the coreassemblage, no grains approaching pure albite have beenfound, and compositions near pure KAlSi3Os are not

510 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS

.9Eoo

o,=rt)

L

or

.9 0.9EIo

I o.tozoz

0.7

0.7

0.6

0.5

0.4

.uEoo

_q 0.8

ozo- o.7

0 .10.0 0.1 0.2 0.3 0.1 0.5

Ti'', ions,/23 O

0.3 0.4 0.5 0.6Fel( Fe+Mg), qfomic

0 . 10.0 0.2 0.3 0.4 0.5Ti"', ions/23 O

o.20.6L

0 l 0.1 0.6 0.8 l .0 1.2F, ions/23 O

0.8

0.7

uEIo

o()+ozoz

o'2 o'4 06 or,t ' . . ] ;7rg'6

t '4 r '6 r '8

Fig. 7. Compositional relations for zoned alkali amphiboles. Curves qualitatively illustratc relations within inclusions of three sizes.(A) Fe/(Fe * Mg) versus TiYIcontent. (B) Na/(Na + Ca) versus Tiu content. (C) Na/(Na + Ca) versus Fe/(Fe + Mg). (D) Fel(Fe + Mg)versus F content. (E) Na/(Na * Ca) versus F content.

A

.cEoo

o)=+q)

L

oL

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS

Table 4. Electron-microprobe analyses and structural formulae (based on 32 oxygen ions) offeldspar grains within lithic-wacke country

country rock

C Y P l O I , A A C\P7 ,D? CYP52 ,G2 cYP8 ,C

Averager Range* Dus ty C lea rCore Rim

511

Core R im Core R im

rock and inclusions

Cores o f la rger' I i t h i c - w a c k e i n c l u s i o n s

s i02

A I203

Ca0

Nar0

K2o

To ta l

siA I

Sum

Ca

Na

K

Sum

An

Ab

0 r

64 .9

1 9 . 3

0 .02

0 .08

I 6 . 3 6

59 .15 67 .6 - 70 ,7

19 .74 19 .0 - 20 .4

0 .04 0 .00 - 0 . 09

1 1 . 6 7 I 1 . 4 - 1 1 . 8

0 .05 0 .02 - 0 , 08

100 .55

11 .990

4.040

66.2 66.7

2 0 . 9 1 9 . 8

0 .80 0 .33

8 . 2 8 6 . 3 7

4 , 2 4 7 . 5 9

68 .3 66 .0

1 9 . 5 1 9 . 5

0 . 1 4 0 . 1 5

9 . 4 7 2 . 1 4

3 . 2 8 1 3 . 8 I

65 . 7 65 , 5

1 9 . 0 1 9 . I

0 . 0 9 0 . 0 7

0 . 1 7 0 . 3 2

16 .28 15 .83

100.42 100.79 100.69 101. s0

11.708 11.868 71.972 1r .892

4 .356 4 .160 4 .036 4 .148

r07.24 100.82 100,66

11 .968 11 .960 11 .908

4 .080 4 .712 4 .154

16.030

0.0r03 .9200,010

3.940

0 . 299.60 . 2

15.054 16.028 16.008

0 .152 0 .064 0 .024

2.840 ?.196 3.220

0 .956 7 .724 0 .732

16.048 16.072 15.072

0 ,020 0 .012 0 .004

0 .060 0 .116 0 .028

3 .784 3 .684 3 .828

16.040

0 .028

0 . 748

3 . t 7 6

3 .948 3 .984 3 ,976

3 . 9 1 . 6 0 . 6

7t .9 55 . I 81 .0

24 .2 43 .3 I 8 .4

3 .864 3 .812 3 .860

0 . 5 0 . 4 0 . 1

1 . 6 3 . 0 0 . 7

97.9 96.6 99,2

3.952

0 . 7

18 ,9

80 .4

*For 14 analyses of I grains.

reached. Quartz and lithic fragments makc uP about 60%of sample CYP4 and typically 55 to 90o/o of Franciscangraywacke (Bailey et al., 1964), yet are raro in the inclusioncoros; it appears, therefore, that a significant proportion ofthe intermediate alkali feldspar in these cores has be€ncreated by feldspathization of these constituents. In con-trast to the cores, the predominant feldspar in more uni-formly metasomatized small inclusions is nearly end-member KAlSi3Or(Ore6.6-ee.2Abo.?-r.oAno.r-o.r). Suchgrains may be "dusty" or clear, or may display a dusty coreand clear rim (Figs. 2C and 2E), showing little compo-sitional contr{rst (Table 4). Within the rim assemblage oflarger inclusions, feldspars show a greater range in compo-sition (we have analyzed a grain core containing 9.4 wt.o/oNarO in the rim of sample CYP101); nearly complete ex-change is typical only near the margin of the inclusion.Within the rim assemblage, feldspar grains are commonlysurrounded by phases (e.g., aonite or natrolite-like zeolite)that presumably have incorporated a part of the Na re-leased by alkali exchange.

Zeolites and related phases

Table 5 lists analyses of the natrolitelike zeolite andrelated phases. Its common occurren@ as an interstitialphase (e.g., Figs. 2C, 2E, and 2F) suggests that thenatroliteJike zeolite crystallized last. Both the natrolitelikezeolite and the fibrous, unidentified Ca-Na silicate aremore conspicuous in the smaller, more nearly uniformly

recrystallized inclusions and commonly are intimately in-

tergrown (Fig. 2F); it is dilficult to ascertain whether thephases are in a reaction relation' Spatial distribution sug-gests that the natrolite-like zeolite, which is the dominantzeolite species toward the edges of smaller inclusions, isencroaching on the fibrous Ca-Na silicate, which is moreabundant toward the centers of these inclusions. However,formation of the less common fibrous silicate may alsohave resulted from local chemical variations' Identificationof the Na-rich zeolite as a natrolitelike mineral is basednot only on the microprobe analyses but also on the X-raypowder-diffraction pattern which is similar to that ofnatrolite.

The X-ray powder-diffraction pattern of the fibrous, un-identified Ca-Na silicate in sample CYP52 is nearly identi-cal to that of gonnardite (XPDF 10-473). However, theel€ctron-microprobe analyses of this phase differ fromthose reported for gonnardite. Perhaps the Ca-Na silicateis a new mineral which has a crystal structure similar tothat of the gonnardite but is substantially different inchemical composition; or perhaps this phase is a gon-

nardite that has undergone alteration and changed compo-sition but still retains enough of the structure to give thegonnardite X-ray pattern. Further study of this phase ismade diffrcult by its intergrown and fine-grained oc-curren@ (Fig. 2F).

Stilpnomelane is a common phase in Franciscan wackes(Ernst, 1965) and Table 6 lists analyses of three grains of a

512 CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIB1LES IN INcTUsIoNs

Table 5. Electron-microprobe analyses ofnatrolitelike and related phases

i b rous un P e c t o l i t e a s t o n i t e

CYP8,A CYP8,B CYP178,A cYP52 ,A CYP52 ,B cYP52 ,C cYPs?,0 C\P52,E cYP7 ,C CYPT ,Fs i02

A',I203

Ca0

Na20

Kzo

Tota l

39 .3

2 7 . 4

? , $

39 . 3 4 3 . 64 3 . 5 47 .2

1 0 . 1 t 2 . 1 1 1 . 3

0 .06 0 .27 0 .05

5 2 . 9 5 1 . 8

0 . 0 0 0 . 0 7

33.8 45 .4

8 . 5 3 0 . 0 5

0 . 0 7 0 .00

44.4

26 .9 26 .L 25 .8 26 .2 t 2 .2 13 .5 15 .1

3 .40 0 .87 0 .40 o .Lz 18 .9 14 .3 t 4 .2

13 .1 1 .2 .4 t 4 .2 15 .6 15 .8

0 .03 0 .02 0 .06 0 .10 0 .04

81.98 82.0? 8?.73 85.40 85. i6 88.46 86.21 85.05 95.30 st .3z*

r A l s o : F e o T , l . 1 8 ; M g O , 0 . 2 5 ; a n d M n o , 0 . 3 9 .

stilpnomelaneJike phase from samples CypT and Cyp101.(A typical stilpnomelane analysis is also listed for compari-son.) This phase is common in other samples, although thegrains are typically too small to attempt microprobe analy-ses.

The best X-ray powder-diffraction patterns that could beobtained show a strong broad line centered on 12.3A and aweak broad line at 7.494 that may belong to this phase.Other stilpnomelane lines that should have been presentwere not detected. An attempt to establish the identity ofthe stilpnomelane-like phase through use of the transmis-sion electron microscope (TEM) was of limited success. Agood l2A repeat was s€en, but the crystal showed a greatdeal of disorder. However, microbeam X-ray fluorescenceanalysis concurrent with the TEM study suggested that thehigh and atypical (for stilpnomelane) Ca contents indicatedby some of the routine microprobe analyses may be due tointergrowth with a Ca-rich phase.

Also listed in Table 6 are analyses of several minor sheet_silicate phases in the lithic-wacke country rock and in the

core of a large inclusion (sample CYP178). None of thesephases show any visible relation to stilpnomelane genesis.

QuartzAlthough quartz is a major constituent of the lithic-

wacke country rock at Coyote Peak, it is a minor phase inthe core assemblage of large inclusions and is absent fromsmaller inclusions and the rims of large inclusions.

Mass transferFor purposes of examining mass transfer between the

ultramafic matrix and included lithic-wacke fragments, wecalculated cationic proportions from the whole-rockchemical analyses (Table 1) based on 160 oxygen ions perunit cell (Barth, 1948); Table 7 lists the results. Assumingthat 160 oxygen ions describe a fixed volume regardless ofcationic composition, the number of cations increasesacross the sequence from 89.2 cations/160 oxygen ions inunaltered lithic wacke to 115 cations/160 oxygen ions inthe uncontaminated ultramaflc host.

cYP101 cyPT CYPT

Table 6. Electron microprobe analyses ofsheet-silicate phases in lithic-wacke country rock and inclusions

B i o t i t e ( ? ) y e l l o w S t i l p n o _Muscov i te Ch lor i te remnant phase St i ' l pnome' lane- l i ke phases -me lane*

CYP4 CYP4 CYP4 CYPI78

s i02T i02A l 2 0 ?r"dT-Mn0

M9o

Ca0

NaZ0

Kzo

F

49 . 1

0 .50

26.24 .47

0 .00

? . 5 6

0 .00

0 .08

r i . 3

0 . 2 0

2 5 . 3

0 . 0 3

2 L . T27 .1.

1 . 2 0

1 3 . 5

0.00

0 . 0 5

0.00

0 .09

5 5 . 6

0 .00

1 . ? 61 7 . 9

0 . 3 2

8 .89

1 . 1 8

2 . 3 4

2 . 2 9

0 .02

49.9

0 .00

0 .581 6 . 4

0 . 1 77 . 94

7 . 3 1

0 .04

0 .00

0 .00

49.9

0 .00

0 . 5 8t 2 . 8

0 . 2 9

1 0 . 0

7 . 2 3

0 .06

0 .00

0 .00

44 .52

0 . 1 0

7 , 7 927 .86

0 . 4 2

f , . o J

0 . 2 3

0 . 3 2

1 . 7 7

32 .3 33 .81 . 0 0 0 . 3 6

1 6 . 9 1 6 . 423.8 24.0

0 ,28 0 .2012 .0 11 .3

0 , 2 3 1 . 0 5

0 . 1 1 0 . 1 87 . 9 2 0 . 0 70 .15 0 .00

To ta l 88 .37 94.47 87.69 87 . 36

*Dee r e t a l . ( 1962 , p . 109 , t ab l e 20 , Ana l ys i s 9 ) .

89.80 82.34 80. 96 88 .04

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUS/ONS 5I3

Table 7. Cations per 160 oxygen ions for lithic-wacke country rock, lithic-wacke inclusions, augite rinds, and ultramafic host

L i t h i c -wacke I nc l us i on (5x8 cm) I nc l us i oncountry rock Core Rim r im

Inc lus i ons(2x3 cm)

Augi te Ul t ranaf icr i nd hos t

CYPIOlA CYPlOlB cYP103 CYP5

S i

A1

Fe"_ ? +f e -

M9

N a

K

T i

P

Mn

B a

Sum

6 1 . 0

1 3 . 4

1 . 0 3

2 . 3 4

U . J O

7 . 2 9

0 .90

0 . 3 9

0 .09

0 .04

0 ,02

5 6 . 5

77 ,6

1 . 2 8

1 . 3 6

t . 7 7

1 . 7 6

7 . 78

9 , 2 0

0 .38

0 .08

0 .05

0 . 0 5

1 7 q

1 . 1 9

I . 1 8

? .06

1 . 61

5 . 3 7

13.2

0 . 3 3

0 .08

0 .04

0 . 0 3

56.7

77.2

I . 40

r . 7 7

2 , 7 1

2 . 7 8

4 .95

t3.2

0 . 37

0 . 1 0

0 .04

0 . 0 1

57 .2

17,0

I . 87

0 . 3 6

1 . 7 5

1 . 4 8

6 , 2 8

1 4 . 1

0 . 30

0 ,02

0 . 0 2

0 . 0 2

5 7 . 4

l D . 5

2 .2 r0 . 66

2 . 5 7

U . J J

6 . 54

14.7

0 . 30

0 . 0 2

0 .04

0 .03

5 3 . 1 3 6 . 6

0 . 6 3 1 2 . 0

1 . 6 0 6 . 4 1a 7 0 q 7 0

18.4 15.2

2 r , 4 22 .7

2 . 8 6 9 . 9 5( . r a

0 . 5 8 2 . 5 5

7 . 8 2

0 . 3 1 0 . ? 8

0 .07

89 .23 97 .81 100.29 i00 .40 106 .67 114 .95

There are exceptionally large contrasts in Si, Ca, Mg,Fe2*, Fe3*, Ti, and P contents between unaltered lithicwacke and the ultramafic host. As Si was lost from thelithic-wacke fragments to the strongly silica-undersaturatedultramafic melt, K "flooded" the inclusions due to its highactivity in that melt. Recall that stable K-bearing phases insample CYP5 (and other samples of the uncontaminatedhost) are phlogopite (average Xptlog, 0.85) and nephelinecontaining about 7 wt.Vo KrO (approximately NerrKs25).Additional indication of high K activity in the ultramafichost is the occurrence of late-stage K-Fe sulfide minerals(Czamanske et al., 1978, L979, L98l). Loss of Si also may belargely responsible for the apparent increase in the cationicconcentration of Al in the inclusions; the resulting averageSi/Al of 3.3 approaches the value of 3.0 characteristic ofalkali feldspar. The monomineralic augite rind that devel-oped along the contacts of inclusions with the ultramaficmagma concentrated Ca, Mgn and Fe2* relative to thelithic wacke, and Mg and Fe2+ relative to the ultramafichost. Growth of this rind was promoted by movement of Sifrom the inclusions outward toward the undersaturatedmatrix,

Metasomatic reaction caused little change in the cationiccontents of Fe, Mg, Ti, Mn, and P within the inclusions,despite large gradients in concentration against the ultra-mafic melt. Ca-cationic content increased slightly, whereasNa content decreased, a relation suggesting that all the Nafrom exchanged albite was not fixed locally in new phases'The steady increase in Fe3+/(Fe3* + Fe2+) with the pro-gression unaltered lithic wackrinclusion cores-inclusionrims-small inclusions (Table 7) is ascribed to coupling ofFe3* with Na+ released from feldspar. Whole-rock datalisted in Table 1 show that HrO was lost and F gainedduring metasomatism of the inclusions. The amphiboledata suggest that F activity was initially high and thendeclined.

Although the flux of components across the in-

clusion/ultramafic melt interface may have been influencedby the intervening augite rind, chemical variation withincoexisting phases in the rim assemblage suggests that thevarious mineral species crystallized in micro-environmentsof rapidly changing chemical potentials and intensive pa-

rameters. Whereas the augite of the rind crystallized with

an essentially fixed Fe/Mg, the sodic pyroxene' in which

substitutions involving Na, Ti, and Fe3+ are important,became depleted in Ti and enriched in the acmite endmember during crystal growth (Table 2). In contrast, early-formed, F-rich richterite evolved to later-crystallizinglow-F magnesio-arfvedsonite, enriched in Ti and contain-ing as much as twice the amount of Fe found in the rich-

terite.A dynamic redistribution of alkalies among coexisting

phases has occurred within the inclusions. Alkali feldsparof subequal Na and K contents in the inclusion cores has

developed from nearly pure albite and evolved' toward the

inclusion/ultramafic interface, first to grains composed ofK-rich rims on cores with Na ev K, and flnally to pure

K-feldspar (microcline) grains. Amphibole crystals havecores in which Na/Ca = I in the M(4) site and rims inwhich Na/Ca: Z in M(4); in sample CYP8, A-site Na is

also more abundant in the rims than in the correspondingcores. Na released from alkali feldspar is also found in

acmite (of fixed maximum Na content) and in the natrolite-like zeolite phase. These coexisting newly formed Na-richphases are undersaturated in Si.

It is diflicult to know over what range of temperature the

redistribution of alkalies took place. The study of Zytianovet al. (1978) indicates that relatively low K/(K + Na) in

aqueous chloride solutions in equilibrium with nephelinecan produce K-rich feldspars. At 0.1 GPa, as temperaturesfall to 400'C, essentially pure microcline will coexist withNe* if K/(K + Na) is greater than about 0'3.

Viewed slightly differently, the replacement of albite bymicrocline plus Na-rich zeolites in a nepheline-bearing

5I4 CZAMANSKE AND ATKIN: TITANIAN A1MITE ALKALI AMPHIBzLES /N INcLUsloNs

system may be related to relative desilication reaction sta_bility. Carmichael et al. (1970) considered the relations of

should precede that of albite.. . " (Carmichael et al., 1970, p.246). Ov observations indicate that in the significantiydifferent environment of diatreme metasomatism, th"i, pro_posal is not substantiated. In comparison with K, Na ismore adaptable in readily entering pyroxenes, amphiboles,and zeolites in substantial concentrations.

DiscussionOur interpretation is that present exposures were within

I km of the surface at the time of dynamic eruption of agas-charged highly fluidized silicate melt. Indications ofauto-brecciation and strong chemical disequilibrium onthin-section and hand-specimen scales are ubiquitous fea-tures of the diatreme, which has nonetheless solidified toform an impressively unaltered, tough, and compact rock.Poorly consolidated fragmental rocks, as well as the HrO-and COr-rich alteration facies of ..typical', ultramafic dia-tremes, are not found.

The challenge in studying the lithic-wacke inclusions is

terite, and arfvedsonite. Schmincke (1969, lg74\,in describ-ing the transition in cooling unit E on Gran Canaria fromwhite to blue-gray (oxidized to reduced), noted a change inthe mafic-mineral assemblage from dominant titaniferousacrnite to "dominant richterite zoned outward to arfved-sonite(?)."

Several excellent experimental and crystal-chemical stud-ies of sodic pyroxenes and alkali amphiboles provide achallenging matrix of information to which our mineral-ogic data can be related. The system at Coyote peak was

alkali amphibole, sodic pyroxene stability extended fartherinto the inclusions. Myriads of fine subhedral to euhedralcrystals (typically smaller than 50 pm) scattered throughoutthe rock remind one of experimental-run products (Figs.28 through 2F). In contrast, the alkali amphibole, whichappears to have crystallized later, occurs as rather widelyseparated larger (maximum 1.25 mm across) crystals (Figs.2D and 2E) within 10 mm of the interface between in-clusion and ultramafic host. Amphibole grains are typicallyfree of inclusions but are commonly intergrown with sodicpyroxene along their margins.

The paragenetic relations of sodic pyroxene and am-

phibole, and the zoning of these phases with respect to Ti,are intriguing problems. In terms of bulk composition(Table l) or cations/160 oxygen ions (Table 7), there is noindication of enrichment in Ti or Na during metasomatism.Most of the arguments of Verhoogen (1962) pertaining tothe Ti content of Mg-Fe pyroxenes are contradicted by ourdata for Na-rich pyroxenes. Figure 4 shows that the tita-nian component of the Coyote peak pyroxene cores maybe the highest yet observed, and a remarkable aspect ofthesodic pyroxenes at Coyote Peak is that their zoning isopposite to that of most known instances of titanianzoning in acmitic pyroxenes. Only Ronsbo et al. (1977)mentioned a few grains of titanian acmite zoned towardcolorless acmite from Tertiary ash layers in Denmark.Consideration of Table 2 shows that Ti content is not sig-nificantly related to any variables other than the con-centration and oxidation state ofFe (Fig. 3).

At least two explanations or a combination thereof maybe oflered for the atypical Ti zonation in the sodic pyrox-enes. It could reflect (1) "capture" of Ti by the first-crystallizing pyroxene or (2) the fact that incorporation ofTia* into sodic pyroxenes is favored by relatively high/rr.The first explanation merits consideration because thepyroxene-stability field is approached from the low-temperature side in the lithic-wacke inclusions, whereas"...the constant association of late-stage titanian-acmiteforming after acmite . . . " (Ferguson , 1977, p.248) is createdin an environment of falling temperature in which the ac-tivity of Tia* may be increasing. However, both this moretypical, late-stage environment and the earliest conditionwithin the inclusions are likely to be characterized by thehighest /o,-hence our second explanation. Both Verhoo-gen (1962) and Flower (1974) emphasize that relatively high

/o, favors incorporation ofTi into pyroxene structures, andnatural occurrences oftitanian acmite can be interpreted toreflect highly oxidizing environments.

Based on the experimental work of Ernst (1962, Fig. ll),we suggest that the apparent precedence of sodic pyroxenecrystallization depended principally on relatively high fo,inherent to the lithic-wacke inclusions. An electrochemi-cally balanced reaction between acmite and richterite.

4NaFe3+Si2o 6 + Caz+ t 4Mg2+ + H2o :

Na(CaNa[MgnFer)SirO22(OH)2 + 2Na+ + 2Fe2 +

* Fe3+ + U2O2 (1)

shows this relation and illustrates that widespread acrnitecrystallization also may have been promoted by the firstflush of Na+ from albite exchange. Although it might beargued that alkaliamphibole crystallization was delayedby relatively late influx of Ca2* and Mg2+ from the ultra-mafic host, we have shown in Table 7 and in our earlierdiscussion that this seems unlikely. Conceivably, the laterand more widely spaced crystallization of amphibole mayalso partly reflect relative ease of crystallization (Gold-smith, 1953).

The excellent experimental studies of Ernst (1962), Hueb-ner and Papike (1970), and Charles (1974, 1975, lg77l,

CZAMANSKE AND ATKIN: TITANIAN ACMITE ALKALI AMPHIBOLES IN INCLUSIONS 5I5

allow discussion of alkali-amphibole stability relations. Thenearly full and relatively constant A-site occupancy in theCoyote Peak amphiboles agrees with their stability at rela-tively low pressure and high temperature (Forbes, 1971;Charles, 1975). Moreover, the chemical environment withinthe inclusion is fully compatible with the statement ofHuebner and Papike (1970, p. 1973) that "The formulafurther suggests that richterite crystallizes in an environ-ment in which the chemical potential of alkalies is greatrelative to that of alumina." (See also Charles, L977, p. 621).

Charles (1977,Fie. 1) showed that richterite of the com-position of the Coyote Peak amphibole cores, approxi-mately MgnFe, is stable at quite low /o, (e.g., 10-21 at600'C and 0.1 GPa on the quartz-fayalite-magnetitebuffer). Ernst (1962, p. 732) found similar stability limits forarfvedsonite and concluded that "Arfvedsonite-bearing as-semblages are presumed to have crystallized at lower oxi-dation states..." Presumably, substitution of Tio* forFe3* in the arfvedsonitic rims of the Coyote Peak am-phiboles may be compatible with even lower/sr. These andother important reactions involving amphiboles and biotitewere thoroughly discussed by Wones (1981).

We envision that the lithic-wacke inclusions reachedsome maximum temperature quickly and then recrystal-lized over a somewhat longer, but still relatively short, timeinterval. As K displaced Na from feldspar and inherent/e,was high due to the presence of detrital hydrous and oxi-dized phases, abundant acmite crystallized. With falling Tand fs", richterite, further stabilized by F, began to formmore widely spaced crystals. Absence of acmite crystalswithin relatively larger, contiguous amphibole grains canbe understood by consideration of reaction (1), which sug-gests that acmite could react readily to form sodic am-phibole. As 7 and /s, continued to decrease, titanianarfvedsonite supplanted richterite as the stable phase, andzoned alkali-amphibole grains were formed.

AcknowledgmentsWe appreciate the assistance of Richard C. Erd and Gordon

Nord with the X-ray diffraction and transmission microscopestudies. Thoughtful reviews by Jane Hammarstrom, Tren Hasel-ton, Stephen Huebner, Malcolm Ross, Donald Voight, and DavidWones helped to clarify our presentation.

ReferencesBailey, E. H., Irwin, W. P., and Jones, D. L. (1964) Franciscan and

related rocks and their sigrrificance in the geology of westernCalifornia. California Division of Mines and Geology, BulletinI 83.

Baker, P. E., Gass, I, G., Harris, P. G., and kMaitre, R. lv. (1964)The volcanological report of the Royal Society expedition toTristan da Cunha. Royal Society, London, Philosophical Trans'actions, 256, Series A,439-578.

Barthi T. F. IV. (1948) Oxygen in rocks: a basis for petrographiccalculations. Journal of Geology, 56, 50-60.

Cameron, M. and Gibbs, G. V. (1973) The crystal structure andbonding of fluor-tremolite: a comparison with hydroxyl tremo-lite. American Mineralogist, 58, 879-888.

Cameron, M. and Papike, J. J. (1981) Structural and chemicalvariations in pyroxenes. American Mineralogist, 66, f-50.

Carmichael, L S. E., Nicholls, J., and Smith, A' L. (1970) Silicaactivity in igneous rocks. American Mineralogist, 55,246-263.

Charles, R. W. (1974) The physical properties of the Mg-Fe rich-terites. American Mineralogist, 59, 518-528.

Charles, R. W. (1975) The phase equilibria of richterite and ferro-richterite. American Mineralogist, ffi, 367 -37 4.

Charles, R. W. (1977) The phase equilibria of intermediate compo'sitions on the pseudobinary NarCaMg.Si.O rr(OHlz'NarCaFerSirO rr(OH\". American Journal of Science, 277' 59+625.

Czamanske, G. K. and Wones, D. R. (1973) Oxidation duringmagmatic differentiation, Finnmarka complex' Oslo area,Norway: Part II, The mafic silicates, Journal of Petrology 14'349-380.

Czamanske, G. K., Meyer, C. E., Erd, R. C., and Norman, M' 8.,ll (1977) The Coyote Peak diatreme, Humboldt County, Cali'fornia (abstr.). EOS, 58, 1247.

Czamanske, G. K., Erd, R. C., Sokolov4 M. N., Dobrovol'skaya,M. G.. and Dmitrieva, M. T. (1979) New data on rasvumite anddjerfi sherite. American Mineralogist, &, 7 7 G7 7 8.

Czamanske, G. K., Leonard, B' F., and Clark, J. R. (1980) Erdite' anew hydrated sodium iron sulfide mineral. American Mineral-ogist, 65, 509-515.

Czamanske, G. K., Erd, R. C., Leonard, B. F., and Clark' J. R.(1981) Bartonite, a new potassium iron sulfide mincral. Ameri-can Mineralogist, 66, 369-375.

Deer, W. A., Howie, R. A., and Zussman, J. (1962) Rock-formingMinerals, Vol. 3, Sheet silicates, Longman, London.

Deer, W. A., Howie, R. A., and Zussman, J. (1978) Rock-formingMinerals, Vol. 2A, Single-chain silicates. John Wiley and Sons,New York.

Erd, R. C. and Czamanske, G. K. (1983) Orickite and coyoteite,two new sulfide minerals from Coyote Peak, HunboldtCountry, California. American Mineralogist, 6E, 245154.

Ernst, W. G. (1962) Synthesis, stability relations, and occurrence of

riebeckite and riebeckite-arfvedsonite solid solutions. Journal ofGeology, 70, 689-736.

Ernst, W. G. (1965) Mineral parageneses in Franciscan metamor-phic rocks, Panoche Pass, California. Geological Society ofAmerica Bulletin, 7 6, 879-914.

Ferguson, A. K. (1977) The natural occurrence of aegirine-neptunite solid solution. Contributions to Mineralogy and Pe-

trology, ffi,247-253.Flower, M. F. J. (1974) Phase relations of titan-acrnite in the

system NarO-FerOr-AlrO.-TiO2-SiO2 at 1000 bars totalwater pressure. American Mineralogist, 59, 53G548-

Forbes, W. C. (1971) Synthesis and stability relations of richterite.American Mineralogist, 56, 997-1004..

Ghose, S. (1965) A scheme of cation distribution in the amphibo-les. Mineralogical Magazine, 35, 46-54.

Goldsmith, J. R. (1953) A "simplexity principle" and its relation to"ease" of crystallization. Journal of Geology, 6l'439451.

Grapes, Rodney, Yagi, Kenzo, and Okumura, Kimio (1979) Aenig-matite, sodic pyroxene, arfvedsonite and associated minerals insyenites from Morotu, Sakhalin. Contributions of Mineralogyand Petrology, 69, 97 -103.

Hawthorne, F. C. (1981) Crystal chemistry of the amphiboles. In

D. R. Veblen, Ed., Amphiboles, Reviews in Mineralogy, vol. 9A'p. l-1O2. Mineralogical Society of America, Washin4on, D. C.

Huebner, J. S. and Papike, J. J. (1970) Synthesis and crystal chem-istry of sodium-potassium richterite (Na$)NaCaMgt

516 CZAMANSKE AND ATKIN: TITANTAN A1MITE ALKALI AMpHIBoLEs IN lrcl,usroNs

SiBO22(OH,F)2: A model for amphiboles. American Mineral_ogist, 55, 1973-1992.

Kitamura, M., Tokonami, M., and Morirnoto, N. (1925) Distri_bution of titanium atoms in oxy-kaersutite. Contributions toMineralogy and Petrology, 51, 167-172.

Leake, B. E. (1978) Nomenclature of amphiboles. Canadian Min-eralogist, 16, 501-520 (Also American Mineralogist, 63, lO23_1052).

Morgan, J. W., Czamanske, G. K., and Wandless, G. A. (l9g5)Origin and evolution of the alkalic ultramafic rocks in theCoyote Peak diatremg Humboldt County, California. Ge-ochimica et Cosmochimica Acta,49, 749-7@.

Nielsen, T. F. D. (1979) The oocurrence and formation of Ti_aegirines in peralkaline syenites. Contributions to Mineralogyand Petrology, 69, 23124/..

Papike, J. J., Ross, M., and Clark, J. R. (1969) Crystal-chemicalcharacterization of clinoamphiboles based on five new structurerefinements. Mineralogical Society of Americ4 Special papr 2,tt7-136.

Papike, J. J., Cameron, K. L, and Baldwin, K. (1974) Amphibolesand pyroxenes: characterization oI other than quadrilateralcomponents and estimates of ferric iron from microprobe data.Geological Society of America, Abstracts with programs 6,1053-1054.

Robinson, Peter (1980) The composition space of terrestrialpyroxenes-internal and external limits. In C. T. prewitt, Ed.,Pyroxenes, Reviews in Mineralogy, vol. 7, p. 4lg4g4. Mineral_ogical Society of America, Washington, D, C.

Ronsbo, J. G., Pedersen, A. K., and Engell, J. (1977) Titan-aegirinefrom early Tertiary ash layers in northern Denmark. Lithos, le193-204.

Rosenberg, P. 8., and Foit, F. F., Jr. (1977) Fe2 *-F avoidance insilicates. Geochimica et Cosmochimica Acta, 41,345-346.

Scheumann, K. H., 1913, Petrographische Untersuchungen anGesteinen des Polzengeibietes in Nord Biihmen; insbesondere

iiber die Spaltungsserie der PolzenirTrachydolerit-phonolith-Reihe, in Der Abhundlungen der mathematisch-physischenKlasse der Kiinigl. Sdchsischen Gesellschaft der Wissenschaft-en: Leipzig B. G. Teubner, v.32,no.7,p.759.

Schmincke, H.-U. (1969) Ignimbrite sequence on Gran Canaria.Bulletin Volcanologique, 33, ll99-l2l1.

Schmincke, H.-U. (19'14) Volcanological aspects of peralkalinesilcic welded ash-flow tufs. Bulletin Volcanologique,39, Sg4-636.

Scott, P. W. (1976) Crystallization trcnds of pyroxenes from thealkaline volcanic rocks of Tenerife, Canary Islands. Mineral-ogical Magazine,,t0, 805-815.

Verhoogen, l. (1962) Dstribution of titanium between silicatesand oxides in igreous rocks. American Journal of Science, 260,2rr-220.

Whittaker, E. J. W. (1949) The structure of Bolivian crocidolite.Acta Crystallographica, 2, 312-317.

Whittaker, E. J. W. (1960) The crystal chemistry of the amphibo-les. Acta Crystallographica, 13, 291-298.

Whittaker, E. J. W. and Muntus, R. (1970) Ionic radii for use ingeochcmistry. Geochimica et Cosmochimica Acta, 34, 945-956.

Wones, D. R. (lg8l) Mafic silicates as indicators of intensive var-ibles in granitic magmas. Mining Geology [Society of MiningGeologists of Japanl 3 1(4), no. 168, l9l-212.

Yakowita H., Myklebust, R. L., and Heinrich, K. F. J. (1973)FRAME: An onJine correction procedure for quantitative elec-tron probe analyses. National Bureau of Standards TechnicalNote 796.

Zyianov, V. N., Perchuk, L. L., and Podlesskii, K. K. (1978)Nepheline-alkali fcldspar equilibria: L Experimental data andthcrmodynamic calculations. Journal of Petrology, lg, l-4.

Manusoipt receiued, September 6, 1983;acceptedfor publication, January 14, 1985.