PERISTERITE UNMIXING IN THE PLAGIOCLASES

AND METAMORPHIC FACIES SERIES

By

WILLIAM L. BROWN

Institut flir Kristallographie und Petrographie

Eidg. Techn. Hochschule, Zlirich [*]

ABSTRACT

The peristerite plagioclases in the compositional range An2 to An17 are reviewed

and it is concluded that they are unmixed into two phascs of composition about

An0_2 and An26±2· No plagioclase in the bulk composition range An17 to An26 has so far

bcen found to be unmixed.

Plagioclase crystals in metamorphic rocks pro ba bly nucleated in a disordered form

(monoclinic or nearly so at the Ab-rich end). At the time of the last rccordcd deforma­

tion, the albite in greenschist. facies rocks was too ordered to dcvelop mechanical

twins, whereas the plagioclase in almandine-amphibolite facies rocks was slightly but

sufficicntly disordered to devclop mechanical twins. From this and crystallographic reasoning it is assumed that unmixing begins in partly disordered erystals and that

ordering continues with the unmixing. Unmixed <'rystals shonld not develop mechanical

twins. The to p of the peristeritc solvus may well he at about !\i'iO o c, and it. is probably

raised by pressure to a slight extent.

In metamorphic rocks of the andalusite-sillimanite type (:\Iiyashiro (l!l6l)), the

composition of the plagioclase increases regularly with increasing metamorphic grade,

while in rocks of the kyanite-sillimanite type and of the Jow and high-pressure inter­

mediate groups, it rna y show a jump in the range a bo ut. An7 to An20• It is considerd that

this jump is due in some way to the peristerite solvus. This jump provides valid grounds

for facies boundaries, but facies boundaries established on other grounds should not be

modified too hastily, beforc work on the other minerals (cspecially the amphibolcs)

has been done.

Introduction The plagioclases of low -grade metamorphic rocks grew or were

rearranged or were recrystallised at relatively low temperatures, and sufficient periods of time were presumably available for them to

[*] Now at Department of Geology, University of Manchester.

[354]

PERISTERITES IN :IJETAMORPHIC ROCKS 355

approach internal equilibrium. The synthesis of such plagioclases in

the laboratory has not so far been achieved, because of the very long

times involved. Thus natural low-temperature plagioclases are of

great importance, for they alone enable us to fill in the lower parts

of the plagioclase phase diagram. Even so portions of this phase

diagram may be completely unattainable to study, because of the

instability, in the presence of other minerals, of some or all of the

plagioclases in low-grade rocks. On the other hand, the petrological

aspects of this instability are very interesting.

Several complications are known to occur in the structural rela­

tions of the natural low-temperature plagioclases, which are not found

at high temperature. The first is the superstructure shown by crystals

in the approximate composition range An25 to An75• Theories have been

advanced to account for the extra X-ray reflections (Chao and Taylor

(1940); Megaw (1960)), but the possible connection with the schiller

exhibited by the labradorites is not understood. The second complica­

tion, the peristerite unmixing, is found in more acid plagioclases.

Such crystals often exhibit a blue schiller, though it is much less

spectacular than that of the labradorites.

A Review of Peristerite Unmixing Peristerite unmixing is generally little known and was only

recently discovered. The total number of crystals which have been examined by X-rays and found to be unmixed is very small (some crystals have been studied by several investigators). Unmixing has been reported in the literature for 26 crystals of which 9 come from pegmatites, 7 from coarsely crystalline presumably igneous rocks, 5 from metamorphic rocks and 5 from unstated rock types. This is neither a representative nor a sufficient sample, but it appears very hard to find crystals of sa y l O% An which are not unmixed.

Peristerite unmixing was first discovered by Laves ( 1951) using

X-ray single-crystal methods. Details of the unmixing for 8 crystals

in the bulk composition range An5�An17 were given, crystals outside

this range being homogeneous (Laves (1954)). The X-ray single­

crystal photographs of peristerites show two sets of spots and Laves

interpreted this to indicate an unmixing of the crystal into two

plagioclase phases with definite orientation relations. Laves deduced

356 W. L. BROWK

from the lattice angles that the compositions of the exsolved phases are close to An0 and An30 to a first approximation. The above range in bulk composition was confirmed by Gay and Smith (1955); they were able to estimate the compositions of the exsolved phases more ac­curately as An3±2 and An23±2. Ribbe ((1960) p. 636) confirmed the range in bulk composition of plagioclases showing unmixing given by Laves (1954), though he found a greater spread in the lattice angle y* of the exsolved phases (Table 6). Brown ((1960) Table l), on the other hand, found much greater constancy in this angle, but established that the range of bulk composition of unmixed plagioclases extends down to An2 or less. It would thus appear that at low temperatures albite(low) can accomodate virtually no An in solid solution. The variation in y* found by Ribbe for the An-rich phase may perhaps indicate a variation in composition.

Schneider ((1957) pp. 253-258) examined an oligoclase and studied its rate of "homogenisation" at three temperatures. He confirmed the continuous change of the lattice angles of both phases found by Laves ( 1954), and he established the gradual disappearance of the reflections of the Ab-rich phase. Both Brown ((1960a) pp. 322-323) and Ribbe

((1960) p. 635) suggested on different grounds that the states described as "homogeneous" by Laves (1954) and by Schneider (1957) were not truly homogeneous. Peristerite crystals must be heated for a

considerably langer period of time to destroy the schiller, than is required to produce the changes in the lattice constants (Ribbe ( 1960) pp. 635-636). This is understandable for local disorder of Al and Si will be much easier to achieve than uniform distribution of Al and Si throughout the crystal which is needed to eliminate the schiller; the latter distribution involves lang-range diffusion (compare Gold­smith (1952)).

The absence of a blue or bluish-white schiller is not proof of the absence of unmixing (Laves, 1954). All crystals with schiller which have been examined with X-ray single-crystal methods have been found to be unmixed (Laves ( 1954), 4 crystals; Ribbe (1960), 3 crystals; Brown (1960b), 3 crystals), but they are restricted in composition [l] to the range An5-An10• The proportions of the two phases are in the ratios 6: l to 2: l. It is possible that in crystals with less than 5% An

[l] Bøggild ((1924) p. 43) has found a slightly larger compositional range for

crystals showing schiller.

PERISTJ<jRITES IN METAMORPHIC ROCKS 357

the exsolved An-rich portions are too small (or too few) to produce visible schiller, whereas in crystals with more than 10% An the por­tions are too large (or too numerous). The unmixed phase can be seen in dark-field and in phase-contrast microscopy (Ribbe and Van Cott, (1962)); it occurs as ellipsoidal blebs scattered throughout the crystals and along twin boundaries.

It is clear that the schiller is due to the differences in refractive index between the two phases, as is the case with the moonstones. The view that it is produced by the selective diffraction of blue light by small particles, analogous to the production of the blue colour of the sky, was advanced among others by Viola ((1901) p. 190); it is sup­ported by the observations of Ribbe and Van Cott (1962) on peristerites, and of Spencer ((1930) p. 319) and Raman, Jayaraman, and Srinivasan (1950) on moonstones. The nature of the colours and other observations speak against the reflection theory of Bøggild (1924). Nevertheless, the size, shape, orientation, distribution and number of the exsolved particles must be such as to give an optical diffraction pattern ap­proximating to a reflection, so as to agree with Bøggild's observation that "In one respect the peristerites differ from all other labrador­izing feldspars viz. in giving much hetter reflections. While, in other feldspars, the reflection of the labradorization is a rather large and indistinct spot, it is here very small, and in some instances the cross­formed signal is directly seen." (Bøggild (1924) p. 40). These observa­tions of Bøggild were verified by the writer with a reflecting gonio­meter. It would be interesting to study the schiller using optical methods analogous to those employed in X-ray diffraction.

The effect of the peristerite unmixing may also be seen in other ways. As was pointed out by Ribbe ((1960) p. 641), the unmixed crystals should have slightly higher refractive indices than homo­geneous grains of the same bulk composition. This will influence the refractive index vs. composition diagrams in the peristerite range, though its exact effect may be masked by bulk inhomogeneity, uncertainty in the composition, sample selection etc. There is little agreement on the positions in the refractive index curves where breaks occur or whether they occur at all; compare for example Chayes ((1952) p. 100), Emmons et al. ((1953) p. 31), J. R. Smith ((1955); (1960) p. 218) and Ribbe ((1960) pp. 640-642). The disagreement in the optical data probably arises from the very slight differences

358 --�--�-� -----

W.L. BROWK

to be measured compared with the accuracy of the measurements. In unmixed peristerites the two phases are intergrown with their

a-axes parallel and all faces in the a-axis zone coincide. They thus have (010) in common; the other axes in this face do not coincide, as the angle {3 for the two phases differs by about 15-16', {3 for the Ab­rich phase being about 116°28' and for the An-rich phase 116°13' (as measured from precession photographs). The [101] axis also differs in position by about 12-13', the angle from the a-axis being larger for the Ab-rich phase. Strain results from this misfit, one effect of which is to make the estimation of the composition of the An-rich phase in the peristerites more difficult (Brown ( 1960 b) p. 338; compare the same effect in moonstones and perthites) ; the best estimate is probably about An26±2•

These angular discrepancies are considerably smaller than those found for moonstones by Laves ((l952a) p. 566) and the misfit is less drastic. In the peristerites both phases are triclinic, so the surfaces of intergrowth will be more difficult to predict than in the moonstones -they will probably be determined by structural continuity and least strain effects[2]. It would seem probable that a surface near to (OSI), the reflection surface found by Bøggild ((1924) pp. 40-42), must be of importance in explaining the schiller; it rna y be the surface of greatest extension of the exsolved phase. The development of lamellae approximately parallel to (OSI) rna y indicate a partial continuation of the unmixing analogous to the development of visible lamellae in perthites (Brown, (1960b) pp. 333-335, 339; see also Bøggild (1924) p. 36).

At the albite end unmixing can easily be seen when the ratio of the Ab-rich to the An-rich phase is 50: l or greater. It is reasonable to suppose that this would also hold in the range An17 to An26• Neverthe­less, no plagiocalse in this bulk composition range so far X-rayed has been unmixed. Plagioclases in the range An12 to An17 show unmixing in the ratio of about l: l, and the more acid the plagioclase, the greater the proportion of the Ab-rich phase ( compare Figure 2 in Brown, l960b). The evidence available would suggest that we are dealing with true unmixing and that a solvus exists giving the variation of

[2] See the discussion on the perthites by ,J. V. Smith (1!)61). When both phases

in the perthites are triclinic, the surface of intergrowth is no longer in the b-axis zone (Laves and Soldatos (1!l62a)).

PERISTERITES IN l\IETAl\fORPHIC ROCKS 359

composition of the exsolved phases as a function of temperature [3]. The solvus must have steep sides for the compositions of the two end­members do not vary much, as judged from the lattice angles. Crystals which would appear to "close" the solvus will be hard to tell from submicroscopically zoned crystals. On the other hand the lattice angles may not give a true indication of the composition (see para­graph above).

The apparent rarity or absence of peristerites in the bulk composi­tion range An17-An26 may perhaps be explained as follows (Brown (1960b) pp. 340-343). Unmixing is a process which occurs after the growth of the crystal. Where one exsolved phase greatly exceeds the other in amount, the latter probably occurs as "blebs" encaged by the first. Where the two phases are present in equal amounts, it is pro bable that neither completely encages the other [4]. If now one of the ex­solved phases is not stable in the presence of other minerals in a rock at low temperature, there will be a tendency for it to be replaced by other minerals which are stable. If the An-rich phase becomes unstable before the Ab-rich one, whether it can easily be replaced or not will depend to a large extent on the bulk composition of the plagioclase, i.e. on the extent to which it is encaged, for the encaged phase will be protected by the other. For plagioclases more basic than about An17, the An-rich phase is probably not fully encaged; reaction may readily occur and the bulk composition of the crystal will change towards that of albite. Reaction will perhaps stop at a composition where the An-rich phase is mainly encaged by the Ab-rich phase (i.e. about An17).

Metamorphism involves both an increase and a decrease in temperature. The above sequence of events takes place on increase of order (at constant or falling temperature) from an originally less­ordered state and should be distinguished from the direct crystallisa-

[3] It is possible that no solvus exists, but that the form of the unmixing is a

"loop", similar to the melting loop of the plagioclases. This seems improbable for it

would require a first order phase transition in albite where the loop meets the albite

side of the diagram. All the evidence points to a continuous and gradual orderfdisorder

transition in albite.

['!] It is possible that one phase is always encaged by the other, the "bars" of the

cage becoming thinner and thinner as the amount of the encaged phase decreases, but

this seems less probable. It is clear from such possibilities that the relation between

thermodynamic and X-ray phases need not be a direct one.

360 W.L. BROWN

tion of fully ordered low-temperature states at constant (or rising) temperature. In this latter case both peristerite end-members could theoretically grow either separately or coarsely intergrown; this has never so far been found. These two cases may be compared with the hypersolvus and subsolvus crystallisation of alkali feldspars in granites (Tuttle and Bowen (1958) p. 129; compare also Tuttle (1952)). The crystallisation of plagioclase under metamorphic conditions will be considered in the next section to o btain some light on these two possibilities.

Crystallisation of Plagioclase under Metamorphic Conditions

Metamorphic processes involve the nucleation and growth of crystals, changes within crystals and reaction between crystals. The question to be considered here is: in what state did the plagioclase crystals grow and what changes occurred to them after growth? A new phase must first be nucleated [5]. The nucleus will probably always be less ordered than the stable form, for it probably did not arise under equilibrium conditions (i.e. infinitesimally slowly). For the same reasons subsequent growth may take place by way of a

disordered form. For the whole crystal, the ratio between the rate of growth and the rate of ordering will be significant; this ratio may vary between very much greater and very much smaller than l.

In the potash feldspars, the temperature of the equilibrium diffusive transition between the monoclinic and triclinic forms appears to be about 500°0 (Goldsmith and Laves (1954) pp. 15-16). Laves ((1950) pp. 554-556) has presented evidence showing that all micro­clines with perfect "microcline-type" twinning grew as a monoclinic phase and that the twinning developed on inversion to a triclinic state. Recently Laves and Soldatos (l962b) have found microcline in which the twinning was not quite perfect, and they discussed the possibility that the crystals grew originally in a state which was not quite monoclinic. However, they gave reasons for concluding that such

[5] It is assumed for simplicity that no plagioclase crystals existed in the rock

before metamorphism. This is not the case in many basic and feldspathic rocks and

relict structures are often found, especially in metamorphosed basic la vas. Same of the

changes undergone by relict crystals may be similar to those which new crystals

undergo after growth.

PERISTEIUTES lN '\IETAMORPHIC ROCKS 361

crystals were also once monoclinic. The evidence presented by Laves would thus suggest that the microcline even of metamorphic rocks of low grade grew metastably as a monoclinic or nearly monoclinic phase and inverted to a triclinic phase on ordering. Further ordering involves an extension of the twin domains and this requires a reversal of the Al/Si positions in adjacent domains. Thus ordering will be slow, even geologically speaking, and partially inverted forms may be encountered even in very old rocks (Christie, 1962), though most microclines in low-grade metamorphic rocks have probably inverted to the maximum microcline state. Thus in the potash feldspars ordering is considerably slower than growth.

Laves ((1960) pp. 283-284) suggested that a monoclinic form of albite, monalbite, is stable at temperatures above about 980°C on the basis of experimental results obtained by MacKenzie ((1952 pp. 334-335). MacKenzie ((1957) p. 508), from syntheses of albite, and McConnell and McKie ((1960), p. 443), from a kinetic analysis of MacKenzie's results, suggest that albite(low) is stable below about 400-450°C, a temperature presumably above or about that pertaining in low-grade metamorphism. It does not necessarily follow that plagio­clases in low-grade metamorphic rocks nucleated in the ordered state, for McKenzie ( ( 1957)Table 3) has shown that the initially synthesised albite crystals were disordered. Indeed, it is possible that these synthetic crystals nucleated and grew initially in a monoclinic form and that they inverted to analbite on quenching. This is perhaps suggested by results of MacKenzie ((1952) pp. 334-335) on powders and of Baskin ((1956) pp. 222-223), Schneider ((1957) pp. 252-257) and Brown ((1960 a) pp. 318-320) on single crystals. It has also been shown that monalbite can exist at room temperature (Brown (1960a) pp. 309-3ll).

No evidence has so far been found to suggest that any plagioclase in metamorphic rocks grew originally as monoclinic crystals [6]. Nevertheless, the possibility exists that albite, but not the more basic plagioclases, may nucleate at low temperatures in a monoclinic state. As the rate of ordering in albite is very much faster than in potash feldspar, the crystal might leave no trace of its monoclinic ancestry

[6] Acid plagioclases, especially if they contain potassium, may grow in a mono­

clinic state in lavas and certain igneous rocks; they show complicated inversion rela­

tions.

:362 W. L. BROWK

( compare the reorganised portions of large microcline crystals which show only either albite or pericline twins�Laves, personal communica­tion). Thus, even though the overall rate of ordering may be much faster than the rate of growth, growth may take place by way of a transient disordered phase. All other evidence suggests that the plagio­clases in low-grade metamorphic rocks were essentially ordered�e. g. twinning.

The idea that albite and pericline twinning may be of secondary, i. e. mechanical, origin dates from the end of the last century; such twins have actually been produced experimentally by the deformation of anorthite at room temperature (Mugge and Heide ( 1931); Starkey and Brown (1963)). The fact that Mugge and Heide failed to produce twins in albite(low) led Laves to advance the theory that the ability to twin depends on the orderjdisorder relations in the plagioclases (Laves (1952b and c)). Mechanical twinning may only occur where the SijAl-0 framework is nearly or exactly topologically monoclinic. In anorthite and other plagioclases with doubled c-axis, this is virtually true and twinning is easy. In albite and the acid plagioclases only the disordered forms have a topologically monoclinic framework. The exact relation between order and twinning is at present under investiga­tion.

For pericline twins the theoretical twin plane, the rhombic section, is an irrational plane and its position, defined by the angle a, varies with the lattice constants of the crystal; a is therefore dependent on the composition and structural state of the crystal and on the tempera­ture and pressure. Its position may thus enable us to reach some con­clusions a bont the state of the crystal during the formation of the twins. The variation of a with composition is well known and the variation with structural state has recently been investigated (J. V. Smith (1958 a)). The influence of pressure is at present fully unknown, while the influence of temperature [7] has been considerd to be negligible (J. V. Smith

[7] J. Y. Smith (HJiiRb) has presented evidence that the influence of temperature

on the obliquity, rp, is considerable. On the assumption that the lattice angles of a

heated crystal change directly towards a monoclinic state, he suggested that tempora­

ture would have little effeet on the position of the rhombic section. The writer has

recently found (unpublished data) that the lattice angles af albite(low) change towards

those of maximum mierocline on rapid heating. At lOOO"C albite(low) has ex*= s·i·o 30' (a change of about l 0 20') and y*=!)] c 08' (a change of about 50').

PERISTERITES IN JHETAliORPHIC IWCKS

(1958 a) p. 919). Recent measurements by the writer have shown that the calculated value of the angle a for a very pure albite(low) changes from + 29°38' at room temperature to + 45° 48' at a bo ut 1000 °C, a change of about 16° (Brown, in preparation); the change for anorthite is expected to be much smaller (see Mugge (1930)). This change in albite(low) is interestingly enough in the opposite sense to and much smaller than that caused by increase of disorder-the value for albite(high) at room temperature is about -·5°. Values of a greater than about 30° would indicate that the twins must have developed in an ordered albite. Such twins should not develop mechanically according to Laves' theory (Laves, 1952 c). Val u es less than 30° depend on both the thermal state and the temperature of development. An examination of Figure 5 in J. V. Smith ((1958 a) p. 923) shows that in the range An0 to An50 most of the recorded values for a fall well within the values for the high and low states; this supports Laves' theory, for acid plagioclases in the low form should not develop twins mechanically [8]. The values of a for acid plagioclases in the range 25-37° from Barth (1928) quoted by Smith were all measured on albite in microcline perthites and probably indicate a recrystallisation at low tem peratures ( compare Laves and Soldatos, 1962 a). U nfortuna tel y. no measurements of a have been made on plagioclases from meta­morphic rocks.

Statistical studies of plagioclase twinning have shown that in general twinning is rare in low-grade metamorphic rocks and that it is supposedly completely restricted to albite, acline and pericline types (Phillips (1930) pp. 247-248; Gorai (1951); Turner (1951)). The problem has been examined more closely by Tobi (1961) for meta­morphic rocks in the greenschist and almandine-amphibolite facies (in the sense of Turner and Verhoogen ((1960) pp. 533-553). To bi studied the relative frequency of twins with (010) as composition plane. In the greenschist facies he found that only twins with (010) as composition plane occurred, namely albite, Carlsbad and albite­Carlsbad. This is in agreement with the findings of Jones (1961) for albite in greenschist facies rocks. In the almandine-amphibolite facies

[8] The extrapolation to metamorphic conditions is questionable for man,,- of the

crystals came from pegmatites and veins. In addition, the same values of a could be

obtained by a change of composition from a m01·e An-rich plagioclaso to a more Ab-rieh

one (compare Laves and Schneider (1956)).

364 W. L. BROWN

on the other hand only 23% of all twins had (O lO) as composition plane (Tobi, p. 579). On the basis of Tobi's results it seems probable that the albite crystals in greenschist facies rocks show only growth twins (Carls bad twins can only develop by growth); thus at the time of the deformation of the rocks the albite was too ordered to develop mechanical twins. In the almandine-amphibolite facies the disorder in the plagioclase crystals, due both to the presumably higher temper­ature and to the presence of An in solid solution, was presumably great enough for pericline and perhaps albite twins to develop by deforma­tion. In addition, it should be remembered that the period of the main deformation need not coincide with the time of highest temperature. and the change of temperature with time need not be simple so that additional complications may arise (see Dietrich (1960) p. 44).

Plagioclase Composition and Metamorphic Grade

The question of the change of plagioclase composition with increasing metamorphic grade and whether it is continuous or not will be discussed in this section in relation to the metamorphic facies series recently introduced by Miyashiro ((1961) pp. 277-287). While using the same facies names as used by Eskola ( 1939), Miyashiro distinguished different metamorphic facies series and named them after some of the characteristic minerals developed in the rocks. On the basis of the aluminosilicates, together with staurolite, jadeite and glaucophane he distinguished three standard types of series and two intermediate groups (pp. 278-283), and correlated them with different physical conditions. In supposed order of increasing rock pressure these are l) the andalusite-sillimanite type (type area, Abukuma Plateau, Japan), 2) the low-pressure intermediate type, where staurolite and andalusite occur, 3) the kyanite-sillimanite type (type area, Grampian Highlands, Scotland), 4) the high-pressure intermediate type, where glaucophane and kyanite occur, and 5) the jadeite­glaucophane type (type-area, Kanto Mountains, Japan). When these characteristic minerals are absent from a particular area, it may still be possible to determine the type from the composition of the minerals of solid solutions series (pp 283-284).

Miyashiro discussed the plagioclases and came to the conclusion that '"plagioclase becomes more calcic with rising temperature in any

PERJSTEHITER IX :tlET.DlOIU'HJC HOC'KS 3fi3

types of metamorphism, but the rate of increase of the Ca content is greater in the andalusite-sillimanite type than in the kyanite-silli­manite type. In the jadeite-glaucophane type, intermediate and calcic plagioclases could not be formed" (p. 284). The change of plagioclase composition [9] with metamorphic grade will be discussed in relation to these five types established by Miyashiro. Only isochem­ical series of rocks will be considered at first. The numbers of the subsections refer to the types of the paragraph above. A considerable amount of petrographic detail must be given to show the complexity of the subject.

l) In the Nakoso district in the Abukuma Plateau, Shid6 ( (l95H) p. 165) found a continuous rise of anorthite con tent with increasing grade in regionally metamorphosed basic rocks. In the lower zone al bite and oligoclase occur with epidote, actinolite and horn blende; in a

higher zone oligoclase, andesine or labradorite occur with hornblende. Andalusite and rarely cordierite occur in pelitic rocks. A similar gradual increase has been found by Miyashiro ( (l95H) pp. 246-247) in basic rocks in the Gosaisyo-Takanuki district, just to the north, except he found a gap in plagioclase composition between An40 and An50 in the middle zone. Plagioclase crystals of this bulk composition occur as zoned crystals, the core being generally An30_40 and the rim An50_65• With increasing grade the two compositional zones become less distinct and then merge. Miyashiro ( l95H) considers that this may be due to a miscibility gap.

In contact metamorphic rocks in the adjacent Iritono district, Shid6 ( 1958) found that epidote and zoisite are missing and that the plagioclase in the lowest zones occurring with actinolite and chlorite is andesine or even labradorite. The high An contents are associated with the absence of epidote (p. 200). Unfortunately she did not give the composition of the plagioclase in the pelitic schists. She concluded (pp. 212-213) that the epidote-amphibolite facies does not exist at low rock pressures, because of the instability of epidote under these conditions ( compare F1gure 5, p. 285 in Miyashiro ( 1961) ).

Compton ( ( 1955) pp. 39-41) measured the composition of plagioclases in metabasaltic rocks from California which had been metamorphosed

[\J] The results and conclusions of this section must be evaluatecl in the light of

possible errors in the determination of the composition of the plagiolasl', of the effect

of disequilibrium, of zoning and of relict Prystals.

:366 W.L. BROWN

by a trondhjemite intrusion; he found a more or less continuous increase in anorthite content towards the batholith. The plagioclases nearest the contact were zoned. Epidote occurs in the outer metamor­phic zone but decreases gradually in amount and is entirely lacking in the inner metamorphic zone. Compton concluded (p. 42) that the assemblage oligoclase-epidote-hornblende of the outer zone replaced the assemblage albite-epidote-hornblende of regional metamorphism because the "stress was low enough in the contact aureole to allow oligoclase to form instead of al bite". He later withdrew this opinion after study of the hornblendes present in the rocks (Compton (1958) p. 905).

2) In the regional metamorphism of greywackes and shales and of basic rocks in Michigan, the conditions probably approached those of the low-pressure intermediate type, for staurolite aften abundantly ac­companied by andalusite, occurs in the greywackes (James (1955) p. 1465). In the basic rocks "the change from the sodic oligoclase of the garnet zone to intermediate or calcic andesine [in the staurolite zone] [lO] is abrupt and doubtless is related to the absence of epidote in the high er zone of metamorphism; in the thin sections studied, no plagioclase was found with a composition in the range An20-An30" (p. 1472).

3) The first explicit evidence known to the writer of the dis­continuous increase in the plagioclase composition with increasing grade in regional metamorphism was found in the greywackes of the Missi series, Manitoba (Ambrose (1936)). The type of the metamorphic series is not exactly known, for rocks only up to the garnet zone are preserved, though it is probable that they belong to the kyanite­sillimanite type in analogy with rocks in Scotland. Ambrose found (p. 268) that "water-clear albite persists with little change through the biotite zone. However, with the appearance of garnet, the anorthite content of the plagioclase rises abruptly form An6_8 to An28_32• A similar rapid conversion of albite to calcic oligoclase or sodic andesine has been noted in rocks of this type by several workers. (Wiseman (1934), p. 385; Phillips (1930) p. 250)". Epidote decreases in amount in the garnet zone in the greywackes.

Kvale ( (1945) pp. 173-176) has shown the scarcity of sodic

[lO] Words in brackets added by writer.

PEIUSTERlTES IN METAJ\fOI{PHIC ROCKS 367

oligoclase in Caledonian rocks in western Norway. He found that epidote occurs with plagioclase up to andesine, though it is not always present in the higher grades (p. 163).

The compositions of plagioclases in epidote-bearing and epidote­free rocks of argillaceous, arenaceous, calcareous and volcanic types from New Hampshire-Vermont have been measured by Lyons ( (1955) pp. 140-141). Only albite occurs in the chlorite zone and in the lower part of the biotite zone, but near the top of the biotite zone oligoclase forms for the first time and becomes common in the garnet zone. There is a jump from An10 to about An23 in the composition of the most calcic plagioclase existing with epidote near the top of the biotite zone. A few rocks in the garnet zone contain sodic oligoclase. The most calcic plagioclase measured occurs in the middle of the garnet zone. Epidote occurring with plagioclase is common up to the top of the garnet zone but the assemblage becomes rare in the kyanite zone; epidote is still found up to the highest part of the kyanite zone in the area.

Fyfe, Turner and Verhoogen ( (1958) p. 218) consider that in general "with rising grade of metamorphism there is a sudden jump from al bite (An0_7) to oligoclase-andesine (An15_30), in quartzo-feldspathic, semipelitic, and basic schists". To avoid difficulties over the determina­tion of the amphiboles, they assign basic rocks with albite to the greenschist facies and those with oligoclase or more calcic plagioclase to the almandine-amphibolite facies.

de Waard (1959) measured the composition of plagioclase in rocks of the kyanite-sillimanite type from Timor and found a very low frequency of plagioclases in the range An5 to An20 in both pelitic and basic schists. He states that "in the field the isopleths of An10 and An20 are generally found very close to each other or are superimposed" (p. 560). The distribution of the plagioclase compositions is such that contours can easily be drawn to give a relatively simple picture of increasing metamorphism. Kyanite occurs in the pelitic schists at high grades and epidote occurs in all rock types in association with plagio­clase up to bytownite (p. 554).

4) Good evidence for a jump in plagioclase composition has been found in the Swiss and Italian Alps. The Pennine and Lepontine Alps belong to the high-pressure intermediate type of metamorphism of Miyashiro ( 1961 ) . Kyanite and sillimanite occur in abundance, but

W. L. BHOWX

alkali-amphiboles are also fairly wide-spread in the low-grade parts (see for example Niggli (1960)). In the Monte Rosa area Bearth (1958) found a jump in composition from albite (with glaucophanitic amphi­bole) to basic oligoclase and andesine (with common hornblende). His results have been extended for the paragenesis calcite-plagioclase by Wenk (1962), who has mapped and contoured the Lepontine Alps on the basis of plagioclase composition. \Venk found a jump between An5 and An20• Wenk also established a jump in plagioclase composi­tion between about An7 and An17 in the carbonate-free paragenesis plagioclase-hornblende-epidote; the contours con form in shape but not cxactly in valne to thosc for calcite-plagioclase [11]. Epidote occurs in the Lepontine Alps along with plagioclase of various compositions up to anorthite (Wenk (1958) p. 494).

5) The significance of the glaucophane-schists is problematical. Ernst (1961) has shown that glaucophane in the presence of H20 alone has a very large P-T field of stability, so that its absence in rocks must be due to the rarity of rocks of suitable composition, or to the require­ment of special physical conditions in rocks of more normal composi­tions or both. Ernst (1959) has, however, shown that glaucophane­schists occur which are virtually identical in chemical composition to greenschists and to epidote-amphibolites. He concluded ( ( 1961) p. 764) that "provided they represent equilibrium assemblages, these glaucophane-schists must have formed under P-T conditions different from those of greenschists and epidote amphibolites of similar che­mical composition."

Miyashiro and Seki (1958a) divided the glaucophane-schist facies into two subfacies, in the first of which epidote, glaucophane and chlorite occur and in the second these three minerals are accompanied by lawsonite and pumpellyite. Though calcite is generally found in such rocks, aragonite may also occur (McKee (1962)). The first sub-

[ll] This non-coincidence of the contours may perhaps be a parallel to the observa­

tion made by \Vhite ((1962) pp. 39-41) that albite occurs in greenschists and An23-An36 in interbedded feldspathic rocks. de \Vaard ((1959) p. 559) also found a difference in

composition in the plagioclases in pelitic and basic schists, once the An20 isopleth has

been passed, those in pelitic schists being richer in An than those in basic schists. White

suggests that " ... the albite-oligoclase transition in the groensehists of South \Vestland

is delayed with respect to that in the quartzo-feldspathic schists. Since all the green­

schists examined contain calcite, this delay may be due to a high partial pressure of

C02". Compare Hamberg ((1945) p. 52).

PERISTERITES IN 1\IELUIORPHIC ROCKS 369

facies differs little from the chlorite zone in Dalradian type meta­morphism, but is placed in the high-pressure intermediate type because of the occurrence of glaucophane (Pennine Alps; Omi district, Japan, Banno (1958); Bessi district, Japan, Miyashiro and Banno (1958)). Higher-grade zones generally belong to the epidote-amphibol­ite and amphibolite facies, though these are not always developed.

The second subfacies occurs in the Kanto Mountains and has been zoned by Seki (1958). Miyashiro and Seki ((1958a) p. 202) believed that jadeite is not stable in this subfacies, but it has been recently discovered in the metamorphic rocks of the Kanto Mountains (see Seki (1960) p. 707). These rocks occur in the lowest-grade part of the area and they are followed by slightly higher-grade rocks belonging to the greenschist facies. Albite is the only plagioclase which occurs in the rocks. The frequent association of glaucophane-schists with eclogitic rocks suggests that the latter could be their high-grade equivalents. This is very uncertain because of the frequent relict nature of the eclogite, though this is not always the case (Bearth (1959) pp. 280-281).

* * *

The above petrologic observations have been restricted to rocks of the same composition as far as possible at different metamorphic grades. The observed changes in the plagioclase composition must, of course, be distinguished from those which occur in rocks of different composition at the same metamorphic grade. This latter effect has been stu di ed in rocks of the kyanite zone of the amphibolite facies in the Lepontine Alps by Wenk (1958). He found that in mica-schists free from epidote, horn­blende, garnet and kyanite the plagioclase is a calcic oligoclase, whereas andesine occurs where these minerals are present [12]. In carbonate­rich rocks only highly basic plagioclases occur. Plagioclase in the range An8 to An15 is restricted to rocks with more than 30% microcline.

One must be careful to distinguish a discontinuity in the rate of increase of the plagioclase composition with increasing metamorphic grade from variations in the frequencies of different compositions in rocks from all grades. In low-grade rocks, only albite occurs. With increasing grade more calcic plagioclases become stable, but the sodic ones still remain intrinsically stable. In high-grade rocks the frequencies of the various plagioclase compositions will depend to a large extent

[12] \Venk ((1936) p. 64) found a similar situation in rocks in Sweden.

NGT- 2±

370 W. L. BROWN

on the frequency of the various bulk compositions. Since rocks virtually free from Ca are very rare, albite and perhaps sodic oligoclase will be rare in high-grade rocks. The fact that they are still intrinsically stable is attested by their occurrence in pegmatites and rocks of granitic composition at the same grade. A diagram showing the frequencies of the various plagioclase compositions in rocks from all grades will have a minimum in the oligoclase range. One may thus only use frequency diagrams to show discontinuities in the rate of increase of composition with metamorphic grade, if rocks of only one composi­tion are included (de Waard (1959) and Noble (1962)).

To summarise it would appear that the rate of increase in the com­position of the most basic plagioclase with increasing metamorphic grade is continuous or nearly so in rocks of the andalusite-sillimanite type and perhaps in the low-pressure intermediate type. A jump occurs between about An7 and about An20 in the kyanite-sillimanite type and the high-pressure intermediate type. In the jadeite-glaucophane type it would seem that only albite exists.

Information on the other important minerals is also summarised here. There is a tendency for epidote to accompany increasingly calcic plagioclases from the andalusite-sillimanite type to the high -pressure intermediate type. In the jadeite-glaucophane type epidote is very common. The relation between garnet and the discontinuity in plagio­clase composition is not clear, for the discontinuity may occur in the upper biotite zone (Lyons (1955) Figure 12, p. 140), at the beginning of the garnet zone (Ambrose (1936) p. 268) or in the garnet zone (Phillips (1930) p. 250; Wiseman (1934) p. 385). A jump also occurs in rocks which only sporadically contain garnet (W enk ( 1962) para­genesis calcite-plagioclase). One other impotant Ca-bearing mineral is hornblende; in general the Na content of hornblende increases with increasing grade in all metamorphic series; at low grade it also in­creases from the andalusite-sillimanite type to the jadeite-glaucophane type (Shido (1958) pp. 204-209; Shido and Miyashiro (1959)). See, however, footnote 18.

Peristerite Unmixing and the Metamorphic Facies Series

The frequent occurrence of saussuritised plagioclases in igneous rocks has led to the conclusion that the An present in the plagioclase

PERISTERITES IN l'iiETA:\IORPHIC ROCKS 371 � - ------------------ -�---

has been replaced by epidote. It has also been assumed that on meta­morphism the epidote occurring with plagioclase is in equilibrium with it and that changes in the composition of the plagioclase will show themselves by an increase or decrease in the amount of epidote [13]. This is an oversimplification of the situation, for the plagioclase will compete with all the other phases in the rock for the components which they contain in common. Equilibrium will be attained under a given set of conditions, when the chemical potentials of the components [14] are the same in all phases.

The first curve showing the relation between plagioclase composi­tion and temperature in the presence of epidote was given by Ramberg ( ( 1945) p p. 51-53; ( 1949) p. 33), though the question had frequently been discussed previously (see also Ramberg (1944) pp. lll-132). Ramberg drew a smooth curve indicating that the composition of the plagioclase increases continuously with increasing temperature. He generalised the situation ((1952) p. 52) to include the effects of other minerals besides epidote [15]. Ramberg's diagrams have been modified to take the subsequent discovery of peristerite unmixing into account by Christie (1959), by Rutland ( (1961) though he drew no curve), and by Noble (1962). Both Christie, and Noble to a certain extent, considered the influence of falling temperature on the unmixing and they thus invoked a further process to get rid of the An-rich exsolved phase (compare Brown (1960b) pp. 340-343). Rutland, on the other hand considered rising temperature and presented evidence for supposing that the plagioclases originally grew in the ordered form. This confusion led to a rediscussion of the situation (Christie (1962), Rutland (1962)), in which the roles of water pressure and disorder were introduced as modifying factors. The situation has been slightly complicated by attempts to use such data to redefine boundaries between metamorphic facies (Rutland (1961 and 1962); Noble (1962)). The complexity of any such transition can be seen from a hypothetical

[13] The composition of epidote is sensitive to the composition of the rock and to

temperature (Rosenqvist (1952) pp. 49 and 60-61; Miyashiro and Seki (1958b)).

[14] It is probably valid to neglect potassium in plagioclases in rocks of low grade,

but this is questionable in rocks of high grade. Sen ((1959) p. 485) found 0.9% Or in

plagioclases from rocks in the amphibolite facies and 4.0% Or in the granulite facies.

[15] Ramberg shows a break in the curves on the basis of the supposed inversion

temperature from albite(high) to albite(low) of 700°C given by Tuttle and Bowen

((1950) p. 574).

372 W.L. BIWWK

reaction given by de Waard (1961), which involves chlorite, actinolite. epidote, quartz, calcite, hornblende and plagioclase.

The stable peristerite solvus will be reconsidered in the light of the above diagrams. If we take MacKenzie's value (1957) for the upper stability limit of albite(low) of about 400-450 oc as correct, then it is probable that the top of the peristerite solvus Iies above this temper­ature. A study of Figure 3 in McConnell and McKie ( (1960) p. 443) indicates that there is a very large change in the equilibrium lattice constants of albite synthesised between about 500 oc and about 600 °C. Taken in conjunction with Figure 7 in Laves ( (1960) p. 280), this would suggest that down to about 600°C (Temperature Tc2 of Laves) the two B sites of the albite contain approximately the same amount of AI. This is probably also the case in the acid plagioclases, though the pre­sence of more than l Al per formula unit will probably lower this temperature slightly (for at the same temperature the presence of the extra Al will give rise to slightly greater disorder than in pure albite). Unmixing possibly takes place then at a temperature slightly below 600°C, when the occupancy of the B sites begins to become appreciably unequal. This temperature is admittedly speculative. This represents a change in the writer's views ( (1960) p. 342).

Mechanical twinning can pro ba bly only develop when the occupancy of the two sets of B sites is about the same, i.e. above the temperature Tc2; this temperature probably decreases with increasing An-content, so that, say, an andesine crystal might still develop mechanical twins at a temperature below that of the top of the solvus.

Pressure will presumably facilitate the unmixing, thus raising the temperature of the top of peristerite solvus. An admittedly crude estimate of the effect of pressure may be obtained by considering measurements for pure albite, on the assumption that the transition albite(low) to albite(high) occurs at one temperature and that the volume change and heat of inversion measured at room temperature may be used at the fictitious inversion temperature. Using values of LI V of 0.38ccjmole (see Ferguson, Traill and Taylor (1958) p. 333), of LIQ of 13 caloriesjgm (Kracek and Neuvonen ( 1952) pp. 314-316; Rosen­qvist (1954) pp. 459-460), and a fictitious temperature of about 900°K, a valne of dTjdP of about 2°C/1000 bars is obtained. A factor of 2 could easily exist in this figure [16]; even so this would only

[16] See p. 374.

i T

Ab Figure l. P-T-X diagram showing the supposedly stable peristerite solvus

(narrowly spaced lines parallel to P axis) cut by the plagioclase breakdown surface (faint curved lines parallel to the T-X plane); isobaric sections are shown at three

pressures P1, P2 and P3• The position of the breakdown surface will be highly dependent

on rock"Composition and the partial pressures of the "volatile" components. The solvus

plunges beneath the breakdown surface at the point S. It has been drawn as a simple

dome for simplicity. Plagioclase is stable above the breakdown surface. The three lines,

a, band c, represent three possible "geothermal" gradients; these sections are shown in

Fig. 2. It is also possible that metamorphism occurs under approximately isobaric

conditions.

The situation at low temperatures and pressures below P 1 may be very complex.

The significance of Shido's observations (1958, p. 200) is not clear, but it may be that

the breakdown surface descends to lower temperatures at lower pressures.

The gap between An40 and An50 found by Miyashiro ((1958) pp. 246-247) probably

represents disequilibrium. It is fairly clear that the intermediate plagioclases do not

represent unmixing in a thermodynamic sense.

Plagioclases are not stable at very high pressures, so the diagram will be cut out at

high pressures. Experimental results are available for albite (Birch and leComte (1960))

but none for anorthite. See Fig. 3c.

374 W. L. BROWN"

indicate a possible rise of about 10�40°C in the temperature of the top of the peristerite solvus in 10 kilobars, corresponding to a depth of about 35 kilometres. The solvus is shown as a function of composi­tion, temperature and pressure in Figure l.

We may assume that surfaces exist giving the breakdown of plagioclase as a function of pressure, temperature, bulk composition and relative humidity of H20 (Thompson (1957) p. 814). One such surface is shown in Figure l. It is assumed to intersect the stable peristerite solvus. Where this surface does not cut the solvus the increase of plagioclase composition with temperature at constant pressure would be continuous; where the sudace cuts the solvus the increase would show a discontinuity. Pressure and temperature may increase together with depth in the crust so the course followed in progressive metamorphism would be a section through this diagram parallel to the composition axis and cutting the PT surface in a line,

a b c

i

2 t3 l 2'

P,T

_jJ l L __ _

Ab An Ab An Ab An

Figure 2. Three sections through Figure l parallel to the composition axis and to

the three lines a, b and c. These sections are essentially similar in form to isobaric

sections. The numbers have the following significance: l, plagioclase; 2, plagioclase

p lus epidote, zoisite, calcite etc.; 2', plagioclase on peristerite solvus p lus other minerals;

3, the other minerals alone; 4, two plagioclases on the peristerite solvus.

So far no case is known where the two plagioclases on the solvus exist in separate

crystals. This may be because they have not been specially sought. On the other hand,

it may be that the unmixing does not proceed far enough, because of strain and kinetic

factors.

[16] Y oder, Stewart and J. R Smith ((1957) p. 208) claim that pressure of H20

raises the alkali feldspar solvus by about 14 °0/1000 bars, a value considerably larger

than the speculative value for the peristerite solvus. Bowen and Tuttle ((1950) pp.

499--500) however claim that pressure has only a slight effect on the alkali feldspar

solvus. See also Orville ((1960) p. 105).

PERISTERITES IN METAMORPHIC ROCKS 375 -----·�����������-·

taken for simplicity as a straight line. Three geologically possible sections are shown in Figure 2.

It is, however, probable that kinetic factors play a large role. Indeed, the differences between the metamorphic facies series may be due to the fact that in some, the plagioclases of the correct composi­tion did not have time to unmix; so in this case no jump would occur. This is perhaps supported by the view of Miyashiro (1958} that the plagioclase composition changes more rapidly relative to changes in the amphiboles in type l than in type 3. Figures l-3 are drawn assuming adequate time and must be interpreted in light of the above considerations.

It is clear that changes in the composition of the rock or in the relativehumidities [17] may change the shape and position of the break­down surface. The following conclusions, however, can perhaps be made: l } the higher the temperature is relative to pressure or the shorter the time at given temperatures, the less likely is a discontinuity in the increase in plagioclase composition, and 2} the lower the relative humidity is relative to temperature, the more likely is a discontinuity (see Rutland (1962) p. 155). In almost all types of metamorphism the relative humidity is sufficiently high relative to the temperature to prevent the existence of An-rich plagioclases in low-grade rocks (see, however, Shido (1958)}.

The published curves which show the composition of plagioclase in equilibrium with epidote or any other mineral must be treated with considerable caution. For rocks under specified variations of bulk composition, temperature, pressure and chemical potentials of the mobile components, such curves indicate the most basic plagioclase in equilibrium with the other minerals. For each change a new curve will exist, thus gi ving rise to a whole band of curves. This is seen in the one case where the plagioclase composition was given as a function of the metamorphic zone (Lyons (1955} Figure 12, p. 140). Lyons' measurements indicate that near the top of the biotite zone, there is a sharp rise in the composition of the most basic plagioclase in rocks which contain epidote, but that rocks with oligoclase and epidote still exist in the garnet zone and even high up in the kyanite zone. Lyons'

[17] Relative humidity should not be confused with high pressure of H20. At igneous contacts at shallow depths the relative huminity may be very high, but the

pressure of H20 will be low.

376 \Y. L. BROWN

results can be generalised for Dalradian type metamorphism to take rock type into account as in Figure 3 b. In other types af metamor­phism and in other areas the diagram will have to be slightly or considerably changed. In the andalusite-sillimanite type no break occurs, whereas in the jadeite-glaucophane type only albite exists and only at low grades. These are also shown in Figure 3.

a b c

l l D c �-i

B i

l l B F

l

t=:---P,T i E A E l - -!

l ----------

Ab An Ab An Ab An

Figure 3. Diagrams giving the composition of the plagioclase in a rock as a func­

tion of metamorphic grade (upwards) and rock type. They are not phase diagrams. The

letters have the following significance: A, no plagioclase; B, plagioclase in basic,

ultra basic and car bona te rocks at middle and high grad es; C, plagioclase in pelitic rocks

under the same conditions; D, plagioclase in quartzo-feldspathic rocks under the same

conditions; E, plagioclase in all plagioclase-bearing rocks at low grade; F, plagioclase

not stable, replaced by jadeitic pyroxene and grossularite-bearing garnet; G, two

plagioclases on the solvus.

a) represents the andalusite-sillimanite type.

b) represents the kyanite-sillimanite type. The bounding cm·ve for pelitic rocks

was drawn below that for basic rocks in the oligoclase-andesine range on the basis of

observations by de Waard ((1959) pp. 559-560) and White ((1962) pp. 39-42). If the top

of the solvus Iies at about 550-600°C, the transition from the greenschist to the alman­

dine-amphibolite facies in the sense of Tumer and Verhoogen (1960) could well lie at

about 500 °C, in agreement with their suggested temperature of 550 oc (pp. 552-553).

c) represents the jadeite-glaucophane type. The field F could possibly represent

"eclogitic" conditions.

In summary it is clear that jumps do occur in the composition of the plagioclases in rock types of varied composition with increasing metamorphic grade in certain metamorphic facies series. It is also possible that the plagioclases steer the changes in the other minerals, e. g. the appearance of almandine garnet, changes in the amphiboles

PERISTERITES IN METAMOR PHIC ROCKS 377

etc. [18]. From a consideration of the petrographic results described in the last section, it seems reasonably clear that, where a jump in plagioclase composition occurs, a facies boundary may be established somewhere in the range of the jump. Other criteria must be used for facies boundaries where no discontinuity occurs in the plagioclase composition.

Acknowledgements The writer is greatly indebted to Professors F. Laves and J. B.

Thompson, Jr. and to Drs. E-an Zen, R.W. R. Rutland and John Starkey for critically reading the manuscript. Responsibility for possible errors or inaccuracies rests with the writer. He has learnt much about the complexities of rocks, compared with the simplicity of theory.

REFERENCES

AMBROSE, J. W. (1936): Progressive kinetic metamorphism in the Missi series near

Flinflon, Manitoba. Amer.Jour. Sci., vol. 32 (1936) 5. ser., pp. 257.

BANNO, S. (1958): Glaucophane schists and associated rocks in the Omi district,

Ntigata Prefecture, Japan. Japanese Jour.Geol. Geography, Trans., vol. 29 (1958) pp. 29.

BARTH, T. F. W. (1928): Die Lage des "rhombischen Schnittes" be i sauren Plagioklasen.

Zeitschr.Krist., vol. 68 (1928) pp. 616.

BASKIN, Y. (1956): Observations on heat-treated authigenic microcline and albite

crystals. Jour.Geol., vol. 64 (1956) pp. 219.

BEARTH, P. (1958): Uber einen Wechsel der Mineralfazies in der Wurzelzone des

Penninikums. Schw.Mineral. Petrogr.Mitt., vol. 38 (1958) pp. 363. --- ( 1959): Uber Eklogite, Glaucophanschiefer und metamorphe Pillowlaven.

Schw.Mineral. Petrogr.Mitt., vol. 39 (1959) pp. 267.

BmcH, F. and LECOMTE, P. (1960): Temperature-pressure plane for albite composition.

Amer.Jour.Sci., vol. 258 (1960) pp. 209.

BøGGILD, O. B. ( 1924) : On the labradorization of the feldspars. Kgl. Danske Videnskab.

Selskab, Mat.-fys.Kl., Medd., vol. 6 (1924) no. 3, pp. 1- 79.

BowEN, N. L. and TUTTLE, O. F. (1950): The system NaA1Si308-KAISi308-H20.

Jour. Geo!., vol. 58 (1950) pp. 489.

BROWN, W. L. (1960a): Lattice changes in heat-treated plagioclases- the existence of

monalbite at room temperature. Zeitschr. Krist., vol. 113 (1960) pp. 297-329.

--- (1960b): The crystallographic and petrologic significance of peristerite un­

mixing in the acid plagioclases. Zeitschr.Krist., vol. 113 (1960) pp. 330.

[18] Sundius ((1946) pp. 32-35) and Shido ((1958) pp. 183-185), however, have

suggested that a miscibility gap exists between actinolite and common hornblende.

See also Compton (1958).

378 �W. L. BROWN

CHAO, S. H. and TAYLOR, W. H. (1940): Isomorphous replacement and superlattice

structures in the plagioclase felspars. Proc.Roy.Soc. (London), Series A, vol.

176 (1940) pp. 76.

CHAYES, F. (1952): Relations between composition and indices of refraction in natura!

plagioclase. Amer.Jour. Sei., Bowen vol., pp. 85.

CHRISTIE, O.H. J. (1959): Note on the equilibrium bctween plagioelasc and epidote.

Norsk Geol.Tidsskr., vol. 39 (1959) pp. 268.

---- (1962): Feldspar structure and the equilibrium between plagioclase and

epidote. Amer.Jour.Sci., vol. 260 (1962) pp. 149.

CoMPTON, R. R. (1955) :Trondhjemite batholith near Bidwell Bar, California. Bull.

Geo!. Soc. Amer., voL66 (1955) pp. 9.

--- (1958): Significance of amphibole paragenesis in the Bidwell Bar region,

California. Amer. Mineral., vol. 43 ( 1958) pp. 890.

DIETRICH, R. V. (1960): Banded gneisses of the Randesund area, southeastern Norway.

Norsk Geo!. Tidsskr., vol. 40 (1960) pp. 13.

EMMONS, R. C. et al. ( 1953): Selected petrogenic rclationships of pla.gioclase. Geo!. Soc.

Am., Mem. 52, pp. 1-142.

Ert::<�ST, W. G. ( 1959): Alkali amphiboles. Annua.! Report, Director Carnegie Inst.vVash.,

58, pp. 121.

---- (1961): Stability relations of glaucophane. Amer.Jour.Sci., vol. 259 (1961)

pp. 735.

EsKOLA, P. in BARTH, T. F. W., CoRRENS, C.W., and ESKOLA, P., (1939): Die Ent­

stehung der Gesteine. Springer V erlag, Berlin.

FERGUSON, R. B., TRAILL, R. J. and TAYLOR, W. H. (1958): The crystal structures

of low-temperature and high-temperature albites. Acta Cryst., vol. ll (1958)

PP· 331. F'YFE, W. S., TURNER, F. J. and VERHOOGEN, J. (!958): Meta.morphic rea.ctions and

meta.morphic fa.cies. Geol.Soc.Amer., Mem. 73 (1958) pp. l. GAY, P and SMITH, ,J. V. (1955): Phase relations in the plagioclase felspa.rs: composi­

tion range An0 to An70• Acta. Cryst., vol. 8 ( 19.55) pp. 64.

GoLDSMITH, J. R. (19.52): Diffusion in pla.gioclase feldspars. Jour. Geo!., vol. 60 (1952)

pp. 288.

--- and LAVES, F. (1954): The microcline-sa.nidine stability relations. Geochim.

Cosmochim. Acta, vol. 5 (1954) pp. l. GonAr, M. (1951): Petrologica.l studies on pla.gioclase twins. Amer.Mineral., vol. 36

(1951) pp. 884.

JAMES, H. L. ( 1955): Zones of regional metamorphism in the Precambrian of northern

Michigan. Bull.Geol.Soc.Amer., vol. 66 (1955) pp. 145i5.

JONES, K. A. ( 1961) : Origin of al bite porphyroblasts in rocks of the Ben More-Am

Binnein area, western Perthshire, Scotland. Geo!. Mag. , vol. 98 (1961) pp. 41.

KRACEK, F. C. and NEUVONEN, K. J. (1952): Thermochemistry of plagioc!ase and

alkali feldspars. Amer.Jour.Sci., Bowen vol., pp. 293.

KvALE, A. (1945): Petrologic and structural studies in the Bcrgsdalen Quadrangle,

western Norway. Part l . Petrography. Bergens Museums Arbok, Naturv. Rekke,

no. l, pp. l .

PERTSTERITES IN METAI\IOR PHIC ROC KS 379

LAVES, F. ( 1950): The lattice and twinning of microcline and other potash feldspars.

Jour. Geol., vol. 58 (1950) pp. 548.

·--- (1951): Relationship between exsolved plagioclase and its host.Amer.Cryst.

Assoc., Abstract Wash. Meet., pp. 33.

---- (1952a): Phase relations of the alkali feldspars. 2. The stable and pseudostable

phase relations in the alkali feldspar system. Jour. Geol., vol. 60 (1952) pp. 549.

---( 1952 b): Uber den Einfluss von Ordnung und Unordnung auf mechanische

Zwillingsbildung. Naturwiss., vol. 39 (1952) p. 546.

---- (l952c): Mechanische Zwillingsbildung in Feldspåten in Abhångigkeit von

Ordnung-Unordnung der Si/A1-Verteilung innerhalb des (Si,Al)408-Geriistes.

Naturwiss., vol. 39 (1952) pp. 546.

---- (1954): The coexistence of two plagioclases in the oligoclase compositional

range. Jour. Geol., vol. 62 (1954) pp. 409.

· --- (1960): AlfSi-Verteilung, Phasen-Transformationen und Namen der Alkali­

feldspåte. Zeitschr.Krist., vol. 113 (1960) pp. 265.

------and SCHNEIDER, T. (1956): Uber den rhombischen Schnitt in sa uren Plagio­

klasen. Schw.Mineral. Petrogr.Mitt., vol. 36 (1956) pp. 622.

- --- and SoLDATos, K. (1962 a): Plate-perthite, a new perthitic intergrowth in

microcline single crystals, a recrystallization product. Zeitschr.Krist., vol. 117

(1962) pp. 209.

- ---- and SOLDATOS, K. (1962 b): Uber "verzerrte" Mikroklin-Verzwilligung und i.\ber

unsymmetrische Albitausscheidung in Kryptoperthit. Zeitschr. Krist., vol. 117

(1962) pp. 218.

LYoNs, J. B. (1955): Geology of the Hanover quadrangle, New Hampshire-Vermont.

Bull. Geo!. Soc.Amer., vol. 66 (1955) pp. 105.

McCoNNELL, ,J. D. C. and McKm, D. (1960): The kinetics of the ordering process in

triclinic NaA1Si308• Mineral. Mag., vol. 32 (1959-61) pp. 436-454.

McKEE, B. (1962): Aragonite in the Franciscan rocks of the Pacheco Pass area, Cali­

fornia. Amer.Minera.I., vol. 47 (1962) pp. 379.

1\fAcKENZIE, \V. S. (1952): The effect of temperature on the symmetry of high­

temperature soda-rich feldspars. Amer.Jour.Sci., Bowen vol., pp. 319.

---- (1957): The crystalline modifications of NaA1Si308• Amer.Jour.Sci., vol. 25/i

( 1957) p p. 481.

MEGAW, HELEN D. (1960): Order and disorder 3. The structure of the intermediate

plagioclase felspars. Proc.Roy.Soc. (London), Series A, vol. 259 (1960) pp. 184.

MrYASHIRO, A. (1958): Regional metamorphism of the Gosaisyo-Takanuki district in

the central Abukuma Plateau. Jour. Fac. Sei. Univ. Tokyo, Sect. 2, vol. ll (1958)

PP· 2HJ.

--- (1961): Evolution of metamorphic beits. Jour. Petrology, vol. 2 (1961) pp. 277.

--- and BANNO, S. (1958): Nature of glaucophanitic metamorphism. Amer. Jour.

Sei., vol. 256 ( 1958) p p. 97.

-----and SEKI, Y. (1958 a): Mineral assemblages and subfacies of the glaucophane­

schist facies. Japan. Jour. Geol. Geography, Trans., vol. 29 (1958) pp. 199.

--------·- and SEKI ( 1958 b): Enlargement of the composition field of epidote and piemont­

ite with rising temperature. Amer.Jour.Sci., vol. 256 (1958) pp. 423.

380 W.L. BIWWN

:\lUGGE, O. (1930): Uber die Lage des rhombischen Schnittes im Anorthit und seine

Benutzung als geologisches Thermometer. Zeitschr.Krist., vol. 75 (1930) pp.

337.

--- and HEIDE, F (1931): Einfache Schiebungen am Anorthit. Neues Jahrb.

Mineral. etc., Beilage Bd.A 64 (1931) pp. 163.

:-.JIGG LI, E. ( 1960): Mineral-Zonen der al pinen l\Ietamorphose in den Schweizer Al pen.

Report 21st Intern. Geoi.Congr.Norden, Part 13, pp. 132.

NoBLE, D. C. (1962): Plagioclase unmixing and the lower boundary of the amphibolite

facies. Jour. Geo!., vol. 70 (1962) pp. 234.

0RVILLE, P.l\1. (1960): Alkali feldspar-alkali chloride hydrothermal ion exchange.

Carnegie Inst. Wash., Year Book 59, pp. 104.

PHILLIPS, F. C. ( 1930): Some mineralogical and chemical changes induced by progressive

metamorphism in the Green Bed group of the Scottish Dalradian. Mineral. Mag.,

vol. 22 ( 1929--31) p p. 239.

R.urAN, C. V., JAYARAMAN, A. and SRINIVASAN, T. K. (1950): The structure and

optical behaviour of the Ceylon moonstones. Proc. Indian Acad. Sei., vol. A 32

(1950) pp. 123.

RAMBERG, H. (1944): En undersøkelse av Vestrandens regionalmetamorfe bergarter.

Norsk Geo!. Tidsskr., vol. 23 (1944) pp. l (with English summary, pp. 153).

--- (1945): Petrological significance of sub-solidus phase transitions in mixed

crystals. Norsk Geoi. Tidsskr., vol. 24 (1945) pp. 42.

--- (1949): The facies classification of rocks: a clue to the origin of quartzo­

feldspathic massifs and veins. Jour. Geo!., vol. 57 (1949) pp. 18.

--- ( 1.952): The origin of metomorphic and metasomatic rocks. Chicago Univ. Press.

HIBBE, P. H. (1960): An X-ray and optical investigation of the peristerite plagioclases.

Am.Mineralogist, vol. 45 (1960) pp. 626.

- --- and VAN COTT, H. C. (1962): Unmixing in peristerite plagioclases observed by

dark-field and phase-contrast microscopy. Canad. Mineral., vol. 7 (1962) pp. 278.

HosENQVIST, I. T. (19.52): The metamorphic facies and the feldspar minerals. Univ.

Bergen Årbok, Naturvitenskap. Rekke, 1952, no. 4, pp. l.

---- (1954): Investigations into the crystal chemistry of feldspars. Proc. Intern.

Symposium Reactivity Solids in Gothenburg 1952, pp. 453-460. Elanders Bok­

tryckeri, Giiteborg 1954.

HuTLAND, R. ,V. R. (1961): The control of anorthite content of plagioclase in meta­

morphic crystallization. Amer.Jour.Sci., vol. 259 (1961) pp. 76.

----- - ( 1962): Feldspar structure and the equilibrium bctween plagioclase and

epidote: a reply. Amer. Jour. Sei., vol. 260 ( 1962) pp. 153.

ScHNEDIDER, T. R. (1957)� Riintgenographische und optische Untersuchung der

Umwandlung Albit-Analbit-Monalbit. Zeitschr.Krist., vol. 109 (1957) pp. 245.

SEKI, Y. (1958): Glaucophanitic regional metamorphism in the Kanto mountains,

central Japan. Japan.Jour. Geol. Geography, Trans., vol. 29 (1958) pp. 233.

--- (1960): Jadeite in Sanbagawa crystalline schists of central Japan. Amer. Jour.

Sei., vol. 258 (1960) pp. 705.

SEN, S. K. (1959): Potassium content of natura! plagioclases and the origin of anti­

perthites. Jour. Geo!., vol. 67 (1959) pp. 479.

PERISTERITES IN l\IETAMORPHIC ROCKS 381

SHmo, FuMIKO (1958): Plutonic and metamorphic rocks of the Nakoso and Iritfmo

districts in the central Abukuma Plateau. Jour. Fac. Sei. Univ. Tokyo, Sect. 2, vol. 11 (1958) pp. 131.

--- and MIYASHIRO, A. (1959): Hornblendes of basic metamorphic rocks.Jour.

Fac. Sci.Univ. Tokyo, Sect. 2, vol. 12 (1959) pp. 85.

SMITH, J.R. (1955): Optical properties of some low-temperature plagioclases. Bull.

Geol. Soc.Amer., vol. 66 (1955) pp. 1618.

--- (1960): Optical properties of low-temperature plagioclases. in Hess, H. H.,

Stillwater igneous complex, Montana. Geol. Soc.Amer., Mem. 80, pp. 191.

SMITH, J. V. (1956): The powder patterns and lattice parameters of plagioclase felspars.

l. The soda-rich plagioclases. Mineral. Mag., vol. 31 (1956-58) pp. 47.

--- (l958a): The effect of composition and structural state on the rhombic section

and pericline twins of plagioclase felspars. Mineral. Mag., vol. 31 (1956-58) pp.

914.

--- (1958 b): The effect of temperature, structural state and composition on the

albite, pericline and acline-A twins of the plagioclase feldspars. Amer.Mineral.,

vol. 43 (1958) pp. 546.

--- (1961): Explanation of strain and orientation effects in perthites. Amer.

Mineralogist, vol. 46 (1961) pp. 1489.

SPENCER, E. (1930): A contribution to the study of moonstones from Ceylon and other

areas, and of the stability-relations of the alkali-felspars. Mineral. Mag., vol. 22 (1929-31) pp. 291.

STARKEY, J. and BROWN, W. L. (1963): Kiinstlicher Erzeugung mechanischer Zwil­

linge in Anorthit. Zeitschr. Krist., in press.

SuNDIUS, N. (1946): The classification of the hornblendes and the solid solution rela­

tions in the amphibole group. Sveriges Geol. Undersiikn., Arsbok 40 (1946)

no. 4, pp. l .

THOMPSON, J. B. (1957): The graphical analysis of mineral assemblages in pelitic schists.

Amer. Mineralogist, vol. 42 (1957) pp. 842.

Tom, A. C. (1961): Pattern of p1 agioclase twinning as a significant rock property.

Koninkl.Ned.Akad.Wetenschap.Proc., vol. B64 (1961) pp. 576.

TuRNER, F. J. (1951): Observations on twinning of plagioclase in metamorphic rocks.

Amer.Mineralogist, vol. 36 (1951) pp. 581.

--- and Verhoogen, J. (1960): Igneous and metamorphic petrology. McGraw-Hill

Book Co., New York.

TuTTLE, O. F. (1952): Origin of the contrasting mineralogy of extrusive and plutonic

salic rocks. Jour. Geol., vol. 60 (1952) pp. 107.

--- and BoWEN, N. L. (1950): High-temperature albite and contiguous feldspars.

Jour.Geol., vol. 58 (1950) pp. 572.

--- and BowEN, N. L. (1958): Origin of granite in the light of experimental studies

in the system NaA1 Si308-KA1 Si308- Si02-H20. Geol.Soc.Amer., Mem. 74,

PP· l . VIOLA, C. (1901): Uber das Glaukisiren verschiedener Feldspathe. Zeitschr.Krist., voL

34 (1901) pp. 171.

DE WAARD, D. (1959): Anorthite content of plagioclase in basic and pelitic schists as

382 W.L.BROWN

related to metamorphic zoning in the Usu massif, Timor.Amer.Jour. Sci., vol.

257 (1959) pp. 553.

--- (1961): Note on coupled reactions in the greenschist-amphibolite transition.

Koninkl. Ned.Akad. Wetenschap. Proc., vol. B64 (1961) pp. 228.

WENK, E. (1936): Zur Genese der Bandergneise von Omo Huvud. Bull. Geo!. Inst. Univ.

Upsala, vol. 26 (1936) pp. 53.

--- (1958): Uber Diskontinuitaten in Plagioklasserien metamorphen Urspungs.

Schw.Mineral.Petrogr.Mitt., vol. 38 (1958) pp. 494.

--- ( 1962): Plagioklas als Indexmineral in den Zentralalpen: Die Paragenese

Calcit-Plagioklas. Schw. Mineral. Petrogr. Mitt., vol. 42 (1962) pp. 139.

WHITE, A. J. R. (1962): Aegirine-riebeckite schists from South Westland, New Zealand.

Jour. Petrology, vol. 3 (1962) pp. 38.

WISEMAN, J. D. H. (1934): The central and south-west Highland epidiorites: a study in

progressive metamorphism. Quart. Jour. Geol. Soc. London, vol. 90 (1934) pp.

354.

YODER, H. S., Stewart, D. B. and SMITH, J. R. (1957): Temary feldspars. Annua!

Report, Director Camegie Inst. Wash., pp. 206.