Ameican Mineralogist, Volume 73, pages 1292-1301, 1988

Prograde versus retrograde chlorite-amphibole intergrowthsin a calc-silicate rock

Trrorras G. Snlnp. PBrrn R. BusncrDepartments of Geology and Chemistry, Arizona State University, Tempe, Arizona 85287, U.S.A.

Ansrnq.cr

Transmission electron microscopy was used to determine structural relationships be-tween chlorite and amphibole and to distinguish prograde from retrograde intergrowths.The prograde reaction chlorite + calcite + qvartz: calcic amphibole + anorthite + CO,+ HrO results in chlorite grains overgrown by amphibole. Prograde chlorite inclusionsoccur with hk} zone axes subparallel to the amphibole c axis but have incoherent andfractured boundaries. These relations are consistent with mimetic growth of amphibole inthe chlorite foliation but not with topotaxial replacement.

Some retrograde chlorite in amphibole occurs with chlorite (001) and amphibole (100)planes parallel and the amphibole c axis parallel to chlorite tl00l, tI t0l. or [I l0]. Coherentboundaries occur parallel to chlorite (001) layers, with no apparent misfit strain. Growthledges of talcJike layers occur along coherent boundaries, showing that chlorite growthcan occur by a single-layer mechanism. Boundaries that are not parallel to chlorite layersare strained such that chlorite layers bend to promote coherence.

Most retrograde intergrowths are associated with fractures and consist of chlorite withminor amounts of talc and unidentified phyllosilicates. These intergrowths vary fromcoherent crystallites to incoherent and poorly crystallized material, reflecting varied con-ditions of alteration. Wide-chain defects are another retrograde alteration product.

INrnonuc:rroN

The metamorphosed argillaceous-carbonate rocks of theWaterville Formation of south-central Maine have beenextensively studied (Ferry, 1976a, 1976b, 1983), and manypossess reaction textures. Sample 18, the subject ofthepresent study, contains prograde intergroWhs where reac-tant chlorite is overgrown by product amphibole, as shownin Plate lB of Ferry Q976a). The purpose of the presentstudy was to investigate orientation and boundary rela-tionships between prograde chlorite inclusions and hostamphibole in order to determine possible reaction rela-tions. The discovery offine-scale retrograde alteration ofamphibole to chlorite, other phyllosilicates, and wide-chain material has allowed a comparison of prograde andretrograde textures within amphibole crystals.

The term "prograde" is used here for reactions thatresult from the addition of heat and that produce assem-blages reflecting equilibrium at the peak temperature ofmetamorphism. Depending on the composition of the in-tergranular fluid, prograde reactions in metacarbonaterocks result in decarbonation and dehydration reactions.Prograde chlorites described here are remnants of mate-rial that was not consumed by the amphibole-formingreaction but that remains as inclusions within amphiboleporphyroblasts.

The term "retrograde" is used here for reactions that

occur with decreasing temperature after the peak of meta-morphism. Many of these reactions involve hydration ofminerals. Because of the low permeability of metamor-phic rocks under low fluid pressure, fracturing is usuallyrequired for fluids to become available for retrograde hy-dration reactions. Minerals that are more hydrous thanthe host amphibole and that occur in fractures or alonggrain boundaries are interpreted as resulting from retro-grade metamorphism. Chlorite in amphibole that is notassociated with fractures is interpreted as remnants ofincomplete prograde reaction.

Because of the small sizes of intergrowths, transmissionelectron microscopy (rnu) is an important technique fordistinguishing prograde from retrograde minerals and forinterpreting reaction mechanisms. rrv studies of alteredorthoamphiboles have revealed a variety of alterationproducts including chlorite, talc, serpentine, and wide-chain biopyriboles (Veblen and Buseck, I 979, I 980, I 98 I ;cressey etal.,1982; whittaker et al., l98l). All of thesereaction products can be coherently intergrown with theorthoamphibole. Retrograde hydration reactions arecommon in rocks that have experienced complex meta-morphic and tectonic histories, and so it is possible tofind minerals that could be ofeither prograde or retrogradeorigin. Distinguishing between the two origins is impor-tant for understanding metamorphic processes and his-tones.

0003-{04xl8 8 / I r | 2-r 29 2S02.OO t292

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS t293

Gror-ocrc sETTTNG

The sample studied is from the Waterville Formation,which belongs to a 5000-m-thick Silurian turbidite se-quence that fills the Merrimack synclinorium in easternNew England. Protolith rocks consisted of graywackesinterbedded with thin layers of quartzite and argillaceouslimestone, with local horizons of pure carbonate rocks(Ferry, 1976a; Osberg, 1979). These sedimentary unitswere deformed, metamorphosed, and intruded by gran-ites during the Acadian orogeny of Devonian age.

The Acadian orogeny compressed and thickened theMerrimack sedimentary unit in three folding events, re-sulting in a polymetamorphic history. The metamor-phism in the Waterville area appears to fit a two-stagemodel similar to the thermal model of Chamberlain andEngland (1985). Initial crustal thickening resulted in largealpine-scale (Fl) folding (Bradley, 1983; Osberg, 1979)and possibly an early metamorphic event correlative withthe initial Acadian metamorphism (Ml) of central Maine(Novak and Holdaway, l98l; Holdaway et al. 1988). Asecond crustal thickening event resulted in the F2 folding(Osberg, 1979) and the peak of regional metamorphism.Intrusion of the Hollowell group of granitic plutons oc-curred at about the same time as the peak of regionalmetamorphism. The combined regional and contact ef-fects appear to correlate with M3 of Holdaway et al.(1988). A third folding event occuffed slmchronously withor slightly after the peak of metamorphism, with no ob-served metamorphic efect. Additional granitic plutonswere intruded in southwestern Maine as late as the Car-boniferous, resulting in M5 prograde and retrogrademetamorphism to the southwest of the Waterville area(Holdaway et al. 1988; Lux and Guidotti, 1985).

Pnrnor,ocv

Metamorphic conditions for the chlorite-amphibole re-action in sample l8 were determined by Ferry Q976a,197 6b, 1983), and the following summary is based on hiswork. The pressure at the peak of metamorphism was3500 + 200 bars. On the basis of the regional tempera-ture trend, the peak temperature reached by sample 18was 460 'C. The metamorphic intergranular fluid wasnearly a binary HrO-CO, mixture, with X.o, = 0.3 duringpeak metamorphism of amphibole-zone rocks.

The decarbonation reaction that produced amphibolefrom chlorite in the Waterville Formation was driven byfluid infiltration. A generalized form of the reaction isgiven by Ferry (1976a) as chlorite + calcite + quartz +intermediate plagioclase : calcic amphibole * calcicplagioclase + water * carbon dioxide; written for end-member compositions, this reaction becomes MgrAlrSir-O,0(OH)8 + 3CaCO. + TSiOr: CarMgrSi.Orr(OH), +CaAlrSirO8 + 3CO, + 3HrO. Mixing of the H,O-richinfiltrating fluid with that produced by the reaction re-sulted in an increase in X.orwith increasing temperature,producing a reaction zone and preservation of reactiontextures.

Amphibole

c=o.53nm

b=1.81 nm

Chlorite

- a=O.54nmz---:-

b=O.93nm

lu"'"u

Fig. l. Schematic illustration ofthe coherent orientation re-lationship between chlorite and amphibole. (a) The c unit vectorof amphibole is parallel to and nearly equal in length to the aunit vector of chlorite, and the D unit vector of amphibole isparallel and approximately twice the length of the b unit vectorofchlorite. (b) Two chlorite layers are nearly equal in thicknessto three amphibole I-beams.

MTxTrul STRUCTURES AND STRUCTURALRELATIONSHIPS

Chlorite and amphibole have structural elements thatallow coherent or semicoherent intergrowths. Klowledgeof their structures helps in understanding the geometricalcontrol in reactions and defects associated with inter-growths.

Chlorite consists of alternating talc-l ike

[Mg,SioO,o(OH),] and brucite-l ike [Mg.(OH)u] sheets(Pauling, 1930). The most common chlorite structure,making up 800/o of all chlorites, is the IIb one-layer poly-type. It has space group C2/m with cell parameters a :

0.54, b : 0.93, c : 1.42 nm, and P : 97" (Bailey andBrown, 1962). This one-layer poly'type commonly has asemirandom stacking sequence, which can be detectedwith electron difraction by streaking parallel to c* alongk + 3n rows in [00], [1l0], or [1l0] zone-axis diffractionpatterns (Spinnler et al., 1984; Iijima and Buseck, 1977).

Amphibole consists of I-beamJike structural units thatare analogous to portions of the talcJike sheets of chlo-rite. Cell parameters for the calcic amphibole from theWaterville Formation were not determined, but they canbe approximated by those of hornblende, which has spacegroup C2/m with cell parameters a: 0.99, b: 1.81, c:0.53 nm, and B : 105" (Papike et al.,1969). Amphibole

Amphibole

b

1294

Fig.2. rcrur micrograph of a large chlorite (Chl) inclusion inamphibole (Amph), with the chlorite layers nearly parallel to theamphibole (100) planes. A small inclusion of anorthite (An) ispresent as well as phyllosilicates (Phy) along the chlorite-anor-thite boundary.

is a member of a group of minerals that comprise thebiopyribole polysomatic series and consists of a sequenceof alternating mica (M) and pyroxene (P) slabs parallel tothe (010) planes (Thompson, 1978). Because the differentbiopyribole minerals consist of the same basic structuralunits, it is possible to combine these units to form newmineral structures.

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS

Although chlorite is not part of the biopyribole poly-somatic series, the talc-like layer of chlorite is a mica unitand can fit coherently with amphibole I-beams. This fitoccurs parallel to the talclike layers, with an orientationrelationship such that the chlorite (001) planes are par-allel to the amphibole (100) planes (Fig. la). Because thetetrahedral sheet ofthe talc-like layer has pseudotrigonalsymmetry, the amphibole I-beams can be coherentlyaligned such that amphibole [001] is parallel to [00],[110], or [1i0] of chlorite. If the chlorite has randomstacking of oneJayer polytypes, these directions appearequivalent in selected-area electron-diffraction (smo)patterns.

For the orientation relationship given above, a semi-coherent boundary along the (010) planes of the twostructures is possible because ofthe coincidence ofclose-packed oxygen planes. The height of two chlorite layers(2c sin P -- 2.84 nm) approximates that of three amphi-bole I-beams (34 sin 0 : 2.86 nm) (Fig. lb). The close-packed planes of oxygen and hydroxyl ions parallel tochlorite (001) and amphibole (100) are nearly coincident,making local coherence possible along other boundaryorientations.

S.INIpI-B PREPARATION AND EXPERIMENTALTECHNIQUES

Thin sections were cut perpendicular to the foliation definedby the chlorite grains in order to maximize the possibility ofviewing along chlorite layers. These sections were studied opti-cally to investigate textural relationships and to choose suitableregions for rru study. Areas were chosen that had amphibolegrains with c axes nearly perpendicular to the plane ofthe thinsection and with overgrowth textures on chlorite. Samples wereion-thinned with 5-kV Ar ions, and removal of amorphous ma-terial was done at I kV. Samples were carbon-coated to preventcharging in the electron beam.

Transmission electron microscopy was performed using aPhilips EM 400T at an accelerating potential of 120 kV. Quali-

Fig. 3. reu micrograph ofa coherent chlorite inclusion in amphibole. The selected-area electron diffraction (snro) pattern showsthat chlorite c* parallels amphibole a*. The array of white spots along the coherent basal boundary represents tunnel structures.Growth ledges (G) occur along the basal boundary.

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS t295

Fig. 4. rru micrograph of a coherently intergrown chlorite (Chl) inclusion at a subgrain boundary (B) within amphibole (Amph).The sero pattern shows that chlorite c* is nearly parallel to amphibole a*. Chlorite layers are slightly bent at the right-hand boundaryto produce coherence. Two extra talclike layers (T) are also present.

tative energy-dispersive X-ray spectroscopy (Eos) analyses wereused for rapid identification of phases. Semiquantitative Eosanalyses were produced using predetermined ft factors on a PDPll/24 comptter with the FoRTRAN program ANEDs, a modifica-tion of a program oeNax written at Arizona State University byP. G. Self. Images of amphibole and wide-chain defects withinamphibole were interpreted in the same manner as those pre-sented by Veblen and Buseck (1981), Whittaker et al. (1981),and Cressey et al. (1982). Images ofchlorite were interpreted interms of layer type and boundary relationships, following thework ofYau et al. (1984) and Veblen and Ferry (1983).

TEM onsnnvATroNs

Transmission electron microscopy was used to deter-mine the structural and morphological relationships be-tween chlorite and host amphibole. Three types of chlo-rite were found in the amphibole: (l) relatively largechlorite grains, several micrometers in length, that do nothave strict orientation relationships or coherent bound-aries with the host amphibole; (2) submicrometer chlo-rites that have coherent boundary relationships; (3) chlo-rites that are associated with fractures and that havevarious structural relationships to the amphibole. Addi-tional alteration reactions produced wide-chain defectsand other phyllosilicates.

Large incoherent chlorites

Chlorite that is observable with the petrographic mi-croscope occurs in many amphibole grains, with (001)cleavage planes nearly parallel either to { I I 0} or to ( I 00)planes of amphibole. The angles between the chlorite [00]or [010] zone axes and the amphibole [001] zone axiswere determined from rEM sample-stage tilts. The anglesvary from l.l to 18.8" and average 5.0o, demonstratingthat the chlorite (001) layers are subparallel to the am-phibole c axis. However, the orientation of these layersrelative to the amphibole (100) varies considerably. Theincoherent relationship between amphibole and the largechlorites is inconsistent with topotaxy, but there is a sub-parallel mimetic relationship.

Boundaries between large chlorite grains and the hostamphibole are commonly fractured (Fig. 2), but there isno indication that this chlorite results from alterationalong fractures. Some of these fractured boundaries arefilled with minerals such as quartz or very fine grainedphyllosilicate (Fig. 2).

Because only the large chlorites are of sumcient size foreos analysis, their compositions are compared to that ofthe chlorite that is not included in amphibole. The averagecomposition of the large chlorite inclusions is (Mgz 64Fe2 or-

r296

Fig. 5. High-magnification micrograph (a) of a growth ledgefrom Fig. 3 and schematic illustrations of (b) a talcJayer ledgeand (c) a bruciteJayer ledge. The spatial distribution of tunnelsin (a) is consistent \ .ith that of the talcJayer ledge model of (b).

Al, ,oTio ,oMno or), nrAl, 3rsi2 6rO,o(OH)*, as determined bysemiquantitative nos. The average composition of non-included chlorite in sample l8 is (MgrrrFe,srAl,3sTioo,-Mno oo), nrAl, 3r Si2 6eOro(OH)., as determined by electronmicroprobe (Ferr),, pers. comm. 1985).

Chlorite coherently intergrown with amphibole

Small chlorite grains that are unrelated to fractures inamphibole but are coherently intergrown with amphibolewere observed only with reu. These chlorites range insize from approximately 50 to 250 nm parallel to thebasal layers and approximately 15 to 30 nm normal tothe basal layers. They are much less common than eitherthe large chlorite grains or fracture-filling material.

The orientation relationship between the coherentchlorite and the host amphibole is as described in theMineral Structure section. The chlorite c* and amphibolea* axes are parallel (Figs. 3 and 4), showing that the chlo-rite (001) layers are parallel to the (100) planes of am-phibole. Streaking parallel to c* precludes distinguishingamong the [100], [1 l0], or 1IIOl ctrtorite zone axes.

Boundaries parallel to chlorite (100) have clearly visi-ble tunnel structures (Fig. 3) that indicate coherence

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS

through the l0- to 20-nm thickness of the sample. Thelength of the coherent boundaries along chlorite (001)relative to lengths ofother boundary orientations suggeststhat growth was most rapid parallel to the chlorite layers.Strained boundaries are indicated by preferential beamdamage (Fig. 3). Boundaries that are not parallel to thechlorite (001) layers show considerable bending of thechlorite layers (Fig. 4).

Growth ledges occur along some chlorite (001) bound-aries (Figs. 3 and 5). The talc-like layer model (Fig. 5b)is consistent with the pattern of tunnels in the micro-graphs, whereas the brucite-like layer model (Fig. 5c) isnot. Single brucitelike layer ledges have not been ob-served.

Chlorite and other phyllosilicates in fractures

Chlorite and other phyllosilicates are common in nar-row fractures cutting amphibole. Three types of fracture-filling materials occur: (1) chlorite that is coherently orsemicoherently intergrown with amphibole, (2) variouslyoriented chlorite, and (3) other poorly crystalline phyl-losilicates that also occur along fractured grain bound-anes.

Fracture-filling chlorite that is semicoherently inter-grown with amphibole (Fig. 6) occurs with approximatelythe same orientation relationship as the coherently inter-grown chlorite described in the previous section. Curva-ture of chlorite layers results in deviations of up to 9.5'from the amphibole (100) planes. Tunnel structures andbent chlorite layers along boundaries (Fig. 6) indicate re-gions of significant coherency. Much of the fracture-fillingmaterial has a packet-like texture consisting of l0- to 50-nm-wide crystallites of chlorite, together with minor talcand mixed-layer material.

Chlorite in cleavage fractures is also oriented so thatits (001) layers parallel the amphibole cleavage planes(Fig. 7). As with coherent and semicoherent intergroWhs,chlorite [100] (or ( I l0)) directions are parallel to amphi-bole [001]. Defects and subgrain textures are abundant insuch intergrowths.

Chlorite and other phyllosilicates are common in frac-tures along inclusion boundaries, indicating that the frac-turing post-dates the formation of the inclusions. Frac-ture-fi l l ing material consists of variously orientedphyllosilicate packets and chrysotilelike material (Fig.8). Because of the small size and the lack of any controlon the orientation of the crystallites, the material is noteasily identified. These intergrowths are the most disor-dered and loosely intergrown of the observed fracture-filling material.

Wide-chain defects

Wide-chain defects occur in the calcic amphibole andconsist primarily of isolated triple chains, but wider-chaindefects and groups of defects are also present. Althoughisolated triple chains could be growth defects, groups as-sociated with fractures (Fig. 6) demonstrate that alterationhas also occurred by a chain-widening mechanism. These

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS 1297

Fig. 6. reu micrograph of chlorite and minor talc (Tlc) that is semicoherently intergrown with amphibole along a fracture. The

SAED pattern shows that c* of chlorite and talc are nearly parallel to amphibole a*. Tunnels (T) are evident along some boundaries.Wide-chain defects (W) extend into the amphibole from the fracture walls.

defects do not have coherent terminations like those ob-served by Veblen and Buseck (1980), but groups com-monly have cooperative terminations (Fig. 9) and appearto form a reaction front for chain-widening alteration.

DrscussrouDistinguishing prograde from retrograde intergrowths

in thin sections is difficult if the features are very small.TEM can provide detailed information on intergro$'thmorphology, crystallinity, orientation, and boundary re-lationships that can be used to make the distinction. Mostprograde and retrograde chlorite can be differentiated bymorphology and by their structural relationships withamphibole. Because these relationships reflect reactionprocesses, they can also be used to interpret prograde andretrograde reaction mechanisms. The variety of retro-

grade products and textures reflects alteration underchanging conditions through the cooling history.

Prograde microstructures

The large chlorites that do not fill fractures in amphi-bole are interpreted as remnants of prograde chlorite,consistent with Ferry's interpretation (1976a) of amphi-bole overgrowths on chlorite. The average chemical com-position of the large incoherent chlorite is similar to thatof nonincluded chlorite determined by Ferry (pers.comm.). This chemical similarity is consistent with, butnot proofof, the prograde interpretation.

The orientations of the large incoherent chlorites areinconsistent with topotaxial amphibole growth. How-ever, the near coincidence of the amphibole [001] zoneaxes and chlorite (001) layers indicates that the amphi-bole is oriented within the foliation defined by the chlo-

t298 SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS

Fig.7. rErvr micrograph of chlorite along amphibole cleavage. The SAED pattem shows that most of the chlorite is oriented parallelto the amphibole (l 10) planes, but other orientations also occur.

rite grains. A possible origin ofthis orientation relation-ship is coherent nucleation of amphibole porphyroblastson chlorite basal surfaces. The porphyroblastic texture ofamphibole grains suggests that nucleation was rate-lim-iting. Coherent nucleation on only a fraction ofthe chlo-rite grains could have resulted in the apparent orientationrelationship. Other subparallel chlorite grains that werenot nucleation sites form the majority of included chlo-rite.

Retrograde microstructures

Fracture-filling chlorite in amphibole is interpreted asthe result of retrograde alteration. Because chlorite wasbeing consumed to form amphibole (Ferry, 1976a, 1983),it is unlikely that chlorite-filled fractures in amphiboleformed during prograde metamorphism.

The small, coherent chlorites are interpreted as retro-grade because of their association with subgrain bound-aries and coherent relationship to amphibole. They ap-pear to be the result of heterogeneous nucleation alongsubgrain boundaries. Growth ofthese chlorites occurs atledges involving (001) layers along coherent (001) bound-anes.

Alteration of amphibole also occurs by growth of wide-chain defects. Isolated wide-chain defects could be either

amphibole gowh defects or alteration products (Veblenand Buseck, 1980). Groups of wide-chain defects repre-sent an alteration mechanism where nucleation occurs onfracture walls or grain boundaries, with growth in the [ 100]and [001] directions. Groups of wide-chain defects in thepresent study are associated with semicoherent chlorite(Fig. 6), suggesting that they are the result of alterationunder similar physical and chemical conditions.

The semicoherent chlorite in fractures appears to bethe result of alteration at lower grade than the coherentchlorite. Like the coherent chlorite, the orientation ofthesemicoherent chlorite is controlled by amphibole and canbe explained by coherent nucleation of chlorite on frac-ture walls and topotaxial replacement of amphibole. Mis-match of the fracture walls suggests significant replace-ment of amphibole. The texture of the semicoherentchlorite, and the presence oftalc packets and mixed lay-ers, is similar to disordered chlorite that formed at 165'C (Ahn and Peacor, 1985); thus the semicoherent chlo-rite may have formed under similar conditions.

The majority of fracture-filling phyllosilicates are notcoherently intergrown and possess many high-energy de-fects indicative of low-temperature conditions. Most con-sist of disordered chlorite that is accompanied by l0- and7-AJayer phyllosilicates. These intergrowths appear to

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS t299

Fig. 8. rEM micrographs of poorly crystalline phyllosilicate in amphibole fractures. (a) Crystallites of 0.7- and 1.0-nm phyllo-

silicates filI an amphibole cleavage fracture, and (b) phyllosilicate packets and chrysotile-like (Ctl) material fill an amphibole fracture.

have grown from multiple nuclei by a variety of growthmechanisms. The most poorly crystalline phyllosilicatematerial represents the lowest-temperature alteration ofthese samples and is similar in texture to illite-smectiteintergrowths in drill-core samples that formed aI 70 'C

lAhn and Peacor, 1985).On the basis of the above interpretation of alteration

grades, slight retrograde alteration of amphibole appearsto have occurred at various times in the cooling history.The relatively high grade alteration could be the result ofmetamorphism associated with late granitic plutons. Pro-

grade effects ofthe Alleghenian Sebago batholith extendas far as 30 km north (Lux and Guidotti, 1985; Holdawayet al., 1988) but are not recognized in the Waterville area(60 km to the northeast). Retrograde metamorphism tothe west of the Waterville area (Novak and Holdaway,1981), as well as the relatively high grade alteration fea-tures observed in the present study, could be distal retro-grade efects of the Sebago batholith. The low-grade al-teration appears to represent interaction with fluids latein the cooling history and probably resulted from inter-action with near-surface waters.

I 300 SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS

Fig. 9. rev micrograph of chain-width defects terminating within amphibole. The terminations are cooperative but not coherent,resulting in strain fields (S) between terminations.

CoNcr-usroNs

In metamorphic rocks, the mineral assemblages andtextural relationships can be a result ofprograde or retro-grade processes. In favorable cases, fine-scale progradeand retrograde intergrowths, such as the chlorite-amphi-bole intergrowths in the present study, can be distin-guished with rnrra.

Relatively large chlorites included in amphibole are in-terpreted as remnants that were not consumed by theamphibole-forming reaction. Growth of amphibole over

chlorite grains was primarily reconstructive, involving in-coherent or semicoherent boundaries. The [001] directionof amphibole is subparallel to (001) layers of chlorite,suggesting that amphibole grew in the foliation of thechlorite grains.

Retrograde chlorite and other phyllosilicates occur insubmicrometer-sized intergrowths that are associated withfractures and subgrain boundaries. Coherent chloritesalong subgrain boundaries are oriented such that chlorite(001) parallels amphibole (100) and chlorite F001, [l 10],or [1-10] parallels amphibole [001]. The fracture-filling

intergrowths range from well-crystallized and semicoher-ent intergrowths with amphibole to poorly crystallizedand randomly oriented intergrowths. Wide-chain defectsindicate another alteration mechanism. The variations inintergrowth textures suggest that alteration occurred atvarious times in the cooling history.

AcrNowr,nocMENTS

We acknowledge John Ferry for supplying samples, electron-microprobeanalyses, and helpful comments. We thank Max Otten, Brad Smith, andSimon Peacock for helpful discussions. We also thank referees Jo Lairdand Don Peacor and an anonymous referee for useful comments on themanuscript. This work was supported by NSF Grants EAR-8408 I 68, andEAR-8708529; microscopy was done at the ASU nnn,r facility, which issupported by the NSF and ASU.

RrrpnnNcns crrnoAhn, J.H., and Peacor, D R (1985) Transmission electron microscopic

study of diagenetic chlorite in Gulf Coast argillaceous sediments Claysand Clay Mineralogy, 33,228-236.

Bailey, S W, and Brom, B.E. (1962) Chlorite polytypism: Regular andsemi-random one-layer structures American Mineralogist, 47, 819-850

Bradley, D.C. (1983) Tectonics ofthe Acadian orogeny in New Englandand adjacent Canada. Journal ofGeology,9l, 381-400

Chamberlain, C D., and England, P.C. (1985) The Acadian thermal his-tory of the Merrimack synclinorium in New Hampshire Journal ofGeology, 93,593-602

Cressey, B A , Whittaker, E.J.W., and Hutchison, J.L (1982) Morphologyand alteration of asbestiform grunerite and anthophyllite Mineralogi-cal Magazine, 46, 77-8'1

Ferry, J.M (1976a) Metamorphism of calcareous sediments in the Water-ville-Vassalboro area, south-central Maine: Mineral reactions andgraphical analysis. American Joumal of Science, 27 6, 841-882

-(1976b) P, T, fcoz and /".o during metamorphism of calcareoussediments in the Waterville-Vassalboro area, south-central Maine.Contributions to Mineralogy and Petrology, 57, ll9-143.

-(1983) Mineral reactions and element migration during metamor-phism ofcalcareous sediments from the Vassalboro formation, south-central Maine American Mineralogist, 68, 334-354.

r30 t

Holdaway, M J., Dutrow, B.L., and Hinton, R W. (1988) Devonian andCarboniferous metamorph:ism in west-central Maine: The muscovite-almandine geobarometer and the staurolile problem revisited. Ameri-can Mineralogist, 73, 2047.

Igima, S., and Buseck, P.R. (1977) Stacking order and disoreder in chlo-rite and mica (abs.). EOS, 58,524-525.

Lux, D R., and Guidotti, C.V (1985) Evidence for extensive Hercynianmetamorphism in western Maine Geology, 13,696-700.

Novak, J.M., and Holdaway, M.J (1981) Metamorphic petrology, min-eral equilibria, and polymetamorphism in the Augusta quadrangle, south-central Maine. American Mineralogist, 66, 5l-69.

Osberg, P.H. (1979) Geologic relationships in south-central Maine. InP.H Osberg and J W Skehan, Eds., The Caledonides in the U S.A., p.

37-62. Weston Observatory of Boston College, Weston, Massachusetts.Papike, J J., Ross, M., and Clark, J R (l 969) Crystal-chemical character-

ization of clinoamphiboles based on five new structure refinements.Mineralogical Society ofAmerica Special Paper 2, ll7-136.

Pauling, L. (1930) The structure ofchlorites. Proceedings ofthe NationalAcademy of Sciences, 16, 57 8-582.

Spinnler, G.E., Sell PG., Iijima, S., and Buseck, P-R. (1984) Stackingdisorder in clinochlore chlorite. American Mineralogist, 69, 252-263.

Thompson, J.B. (1978) Biopyriboles and polysomatic series. AmericanMineralogist, 63, 239-249.

Veblen, D.R , and Buseck, P.R. (1979) Serpentine minerals: Intergrowthsand new combination structures. Science, 206, 1398-1400

-(1980) Microstructures and reaction mechanisms in biopyribolesAmerican Mineralogist, 65, 599-523.

- (1 98 l) Hydrous pyriboles and sheet silicates in pyroxenes and ur-alites: Intergrowth microstructures and reaction mechanisms. Ameri-can Mineralogist, 66, l l07 -1 134.

Veblen, D R, and Ferry, J M. (1983) A TEM sludy of the biotite-chloritereaction and comparison with petrologic observations American Min-eralogist, 66, 1107-1134.

Whittaker, EJ.W., Cressey, B.A, and Hutchison, J.L. (1981) Termina-tions of multiple-chain lamellae in grunerite asbestos. MineralogicalMagazine, 44,27-35

Yau, Y.C., Anovitz, L.M, Essene, E.J., and Peacor, D.R- (1984) Phlog-opite-chlorite reaction mechanisms and physical conditions during ret-rograde reactions in the Marble Formation, Franklin, New Jersey. Con-tributions to Mineralogy and Petrology, 88, 299-306.

MeNuscnrgr RECETVED Noverrrnrn 25, 1987MeNuscnrpr AcCEFTED Aucusr l, 1988

SHARP AND BUSECK: CHLORITE-AMPHIBOLE INTERGROWTHS