46595

 

This is to certify that the
thesis entitled

CONSTITUTIONAL SUPERCOOLING, A MECHANISM
FOR OSCILLATORY ZONING IN PLAGIOCLASE
presented by

Vivian Kay Bust

has been accepted towards fulfillment
of the requirements for

Master's degeena Geologx

Ofim

 

we, mm / /
Date /8/ / (a

0-7639

0V FINES:
25¢ per day per it.

RETUMIK; LIBRARY “MATERIALS:
Place in bookm

mto move
charge fro. circulation records

 

 

 

CONSTITUTIONAL SUPERCOOLING, A MECHANISM
FOR OSCILLATORY ZONING IN PLAGIOCLASE

By

Vivian Kay Bust

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Geology

1980

ABSTRACT

CONSTITUTIONAL SUPERCOOLING, A MECHANISM
FOR OSCILLATORY ZONING IN PLAGIOCLASE

By

Vivian Kay Bust

This study provides evidence which is interpreted to support
constitutional supercooling as a viable mechanism for oscillatory
zoning in plagioclase.

Constitutional supercooling requires concentration gradients
in the liquid immediately adjacent to a growing crystal, therefore,
the presence or absence of these gradients provides the test of the
model. Concentration gradients occur in the glass matrix adjacent
to only those crystals which exhibit oscillatory zoning at the
crystal perimeter. No concentration gradients occur in the glass
matrix adjacent to normally zoned crystallographic faces.

Constitutional supercooling is controlled only by the envir-
onment immediately adjacent to the growing crystal face; therefore,
if this mechanism is valid different zone patterns may occur on dif-
ferent crystal faces of the same crystal. Furthermore, correlation
of oscillatory zoning between crystals should be limited. Two
crystals observed in this study contain oscillatory zoning isolated
to just one crystal face, while the remaining faces are normally
zoned. Also, there appeared to be only limited zone pattern corre-

lation between crystals sampled from the same rock type.

Dedicated to my parents

ii

ACKNOWLEDGMENTS

I would like to express my sincere appreciation to Tom Vogel

for his patience throughout this study.

A special thanks to my friends who encouraged the completion

of this project.
My most heartfelt thanks to my family for their continued

support and understanding.

iii

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES

INTRODUCTION

Theory.
Previous Work

TEST OF MODEL .
ANALYTICAL METHODS

Grain Selection .
Procedure .

ANALYSIS OF DATA .

Compositional Gradients
Quartz Latite
Rhyodacite
Basalt . .

Pattern of Oscillatory Zones
Quartz Latite . . .
Rhyodacite and Basalt

DISCUSSION .
CONCLUSION .
APPENDIX

BIBLIOGRAPHY

PLATES

iv

Page

vi

—-1

 

LIST OF TABLES

Table Page

l Effective Crystal Width and Glass-Crystal Boundary

Types . . . 3l

2 Periphery Zoning Type and Glass Matrix Concentration
Gradients . . . . . . . . . . . . . . . 33

Figure

\OCDVO‘UT

10.
11.
12.
13.
14.
15.
16.
17.
18.
19.

LIST OF FIGURES

An-Ab Binary Phase Diagram

Skematic Profile of Crystal- -Glass Boundary Types and
Chemical Variation at Boundaries .

Quartz Latite Crystal 1
Quartz Latite Crystal 2
Quartz Latite Crystal 3
Quartz Latite Crystal 4
Quartz Latite Crystal 5

Rhyodacite
Rhyodacite
Rhyodacite
Rhyodacite
Rhyodacite
Rhyodacite
Rhyodacite
Rhyodacite

Crystal
Crystal
Crystal
Crystal
Crystal
Crystal
Crystal
Crystal

1

mVOWUTkOON

Quartz Latite 2-ab .

Quartz Latite 4-ab .

Quartz Latite 5-lm .

Quartz Latite 6-ab .

vi

Page

17
20
21
21
22
23
26
26
27
27
27
28
28
29
38
39
40
41

Figure
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.

Rhyodacite 2-pq .
Rhyodacite 3-ab .
Rhyodacite 4-tu .
Rhyodacite 4-ut .
Rhyodacite 4-sr .
Rhyodacite 8-cg .

Basalt l-xo
Basalt l-pz
Basalt 2-kj
Basalt 3-gb

vii

Page
42
42
43
43
44
44
45
45
46
46

INTRODUCTION

Theory
Sibley et a1. (1976) proposed that the periodic oscillatory

zoning present in plagioclase is due to supersaturation of the
liquid which causes concentration gradients at the melt-crystal
interface. This is defined as constitutional supercooling by Rutter
and Chalmers, l953; Chalmers, l964. The purpose of this study is

to test this constitutional supercooling model as a mechanism for
oscillatory zoning in plagioclase.

Constitutional supercooling is a result of solute enrich-
ment in the liquid in contact with an advancing solid-liquid inter-
face having a composition different from that of the bulk liquid.
The liquid in contact with this interface has a lower liquidus
temperature than the bulk liquid further from the interface.

In constitutional supercooling, there is a mutual variation
of solute concentration and liquidus temperature with distance from
the interface. At the interface the solute concentration is maximum
with respect to the bulk liquid and the liquidus temperature is
minimum with respect to the bulk liquid (Chalmers, 1964).

According to Chalmers (1964), under assumed initial steady-
state condition, if the interface remains planar it should be pos-
sible to supercool the liquid by an amount equal to the interval

between the liquidus and solidus. This situation would imply that

the liquid ahead of the interface could be constitutionally
supercooled. This degree of supercooling is never fully realized
because only a small amount of supercooling is sufficient to set

up an instability* which leads to a departure from the steady-state
condition.

Klein and Uhlmann (1974) studied the crystallization behavior
of anorthite from its melt over a range of undercoolings. It was
found that the morphology of the crystal-liquid interface was
faceted, growth took place preferentially in the crystallographic C
direction and that the fraction of preferred growth sites on the
crystal-liquid surface increases with increasing undercooling.

Klein and Ulhmann (1974) suggested that anorthite grew by
the formation and lateral propagation of two-dimensional nuclei on
a planar interface. This situation implies that a certain amount
of undercooling is necessary to nucleate steps on the interface.

To account for the periodic oscillatory zoning in plagio-
clase, Sibley et a1. (1976) proposed a model based on constitutional
supercooling. In this model, the driving force for oscillatory
zoning is the variable growth rate which is produced by varying
degrees of supercooling and the planar nature of the plagioclase
crystal-liquid interface. Oscillatory zoning in this model is
independent of external variables in the magma chamber, i.e.,

confining pressure, hydrostatic pressure and temperature.

 

*Instability is due to interface attachment kinetics.

The mechanism, proposed by Sibley et al. (1976), by which
constitutional supercooling produces oscillatory zoning in plagio-
clase is as follows, with reference to Figure 1. In order to allow
crystallization of a stable nuclei a liquid, of some composition X,
is initially supercooled to temperature T1 by some mechanism other
than constitutional supercooling. The solid will crystallize such
that, at equilibrium, the composition of the solid will be S1 and
the composition of the liquid in contact with the solid will be L1.
During initial growth, if diffusion of the albite molecule away
from the interface or the anorthite molecule to the interface is
less than the growth rate of the crystal,a concentration gradient
will develop in the melt away from the crystal. This situation
causes the formation of a boundary layer adjacent to the growing
crystal which has a composition different from that of the bulk
liquid. It is this boundary layer and not the bulk liquid that
will be of composition L1 when the crystal composition is S].

Assuming the amount of supercooling at the interface is
maintained then after a period of initial cyrstallization, the bulk
liquid composition will be of some intermediate composition (L3)
between the initial liquid composition X and the boundary composi-
tion L]. At this point the flux of solute is maximum because of
the large compositional difference between L1 and L3. As diffusion
of the solute from the bulk liquid to the boundary proceeds the
composition of the boundary layer (L1) becomes more calcic and

migrates off the liquidus curve to some composition L2.

Because plagioclase grows with a faceted, planar interface,
nucleation is impeded and diffusion of the solute through the melt
may cause the liquid to migrate a considerable distance off the
liquidus before the crystal starts to grow again. If plagioclase
grew with a diffuse interface, growth rates would respond immedi-
ately to the changing composition (L2).

The new boundary layer composition L2 represents the super-
cooling necessary to nucleate new steps on the crystal-liquid
interface to begin a new growth cycle. The exact composition of
the new zone cannot be predicted but would be located to the left
of S]. The solid and liquid compositions would then migrate back
toward S1 and L]. The growth rate will exceed the diffusion rate
until the solid and boundary layer composition are at 51 and L].

At 51 and L], the diffusion rates again exceed growth rates and
the cycle will begin to repeat itself.

For each cycle, the maximum anorthite content of a zone
is determined by the position of L2 which is a measure of consti-
tutional supercooling necessary to initiate growth on a planar
interface. The maximum albite composition is determined by the

position of SI and L1 on the binary phase diagram (Figure 1).

Previous Work

 

Bottinga et al. (1966) divided theories which produce
zoning in plagioclase into two categories: those in which oscil-
latory zoning is caused by repeated changes in the plagioclase-

liquid equilibrium variables and the theory of Harloff (1927)

Figure l. An-Ab Binary Phase Diagram.

 

 

 

 

 

 

 

 

meaumcm

Ab

An

FIGURE 1

which is a diffusion-supersaturation model. Vance (1962) and Smith
(1974) also reviewed and summarized theories which may produce oscil-
latory zoning in plagioclase. A modified version of the theory of
Harloff (1927) was presented by Bottinga et al. (1966) to explain
periodic oscillatory zoning in plagioclase.

Bottinga et al. (1966) further developed Harloff's model
and suggested that once the driving force (supersaturation) has
exceeded the minimum required for two dimensional nucleation of
steps on the crystal surface, growth of a zone begins and proceeds
by lateral propagation of steps across the smooth (planar) inter-
face. Relying on work done by Cahn (1960) Bottinga et al. (1966)
reasoned that because two dimensional nucleation of an incomplete
surface (diffuse) is not much more difficult than on a smooth
surface the interface becomes diffuse.

The change from a planar_to a diffuse interface is attrac-
tive because it increases the number of favorable nucleation sites
on the growth surface and a lower driving force is needed to main-
tain a diffuse interface. According to Cahn (1960) at this stage
the interface is propagated normal to itself rather than repeated
by lateral steps across the surface.

The interface remains diffuse as the driving force decreases
and further growth is limited to only the most favorable sites on
the interface. Finally, the interface becomes planar and a steady-
state is reached when growth and diffusion rates are equalized.

The growth cycle will begin once more when supersaturation of the

melt adjacent to the planar interface exceeds the minimum required
for two dimensional nucleation of steps across the crystal surface.

Bottinga's data, which he cited as support of this
diffusion-supersaturation model, was obtained from an electron-
microprobe analysis of the glass-bytownite interface. A "represen-
tative" chemical profile of these analyses, recorded at selected
points along a 40 micron traverse to the interface, is shown in
Figure 3 of Bottinga et al., 1966. This profile shows the presence
of Si, Al, Mg and Fe concentration gradients and the absence of Ca
and Na concentration gradients in the glass adjacent to a bytownite
crystal, which displays faint oscillatory zoning.

Klein and Uhlmann (1974) studied the crystallization
behavior of anorthite over a wide range of undercoolings. Their
data supports growth of plagioclase by two dimensional nucleation
and lateral propagation of steps across a plaggr_interface at all
undercoolings rather than a diffuse interface. In light of Klein
and Uhlmann's work Bottinga's model is suspect because it depends
on a diffuse interface.

Lofgren (1974a) proposed a model for reverse zoning in
plagioclase which is similar to the constitutional supercooling
model proposed by Sibley et al. (1976). However, Lofgren's model
is dependent on both a diffuse interface and the development of

steady-state conditions.

TEST OF MODEL

If constitutional supercooling is a valid process, concen-
tration gradients should occur in the liquid surrounding a growing
crystal which contains oscillatory zoning and the oscillatory zone
patterns in adjacent crystals should be somewhat independent. The
presence of Si, Ca and Na concentration gradients in the glass
matrix adjacent to the oscillatory zoned plagioclase crystals would
be a major test of the model.

Second, if oscillatory zone patterns of each plagioclase
crystal within the same rock sample are independent of each other
then one would expect that a comparison of zone patterns between
plagioclase crystals to yield an insignificant correlation.
Insignificant correlation of oscillatory zone patterns would pro-
vide subsequent support of the model. Niebe (1968) has shown that
only abrupt zoning discontinuities in plagioclase can be correlated
and that the fine oscillatory zoning cannot be correlated.

Because Klein and Uhlmann (1974) have shown that anorthite
grows with a planar interface and an anisotropic manner, one may
expect different oscillatory zone patterns on various crystal faces.
Constitutional supercooling would be supported if different oscil-
latory zoning patterns occur on different crystal faces from the

same grain. This would indicate that both the interface attachment

10

kinetics and the boundary layer melt are the controlling parameters
for oscillatory zoning in plagioclase.

To test the constitutional supercooling model as a mechanism
for oscillatory zoning in plagioclase, the rock must satisfy two
requirements: (1) the plagioclase phenocrysts must have crystallized
from the same magma, i.e., they cannot be xenocrysts; and (2) these
phenocrysts must also be set in a glass matrix. The glass matrix
may be assumed to represent the last liquid surrounding the crystal
at the time the magma was quenched.

The rock types selected to meet the above requirements are
recent volcanics containing normal and oscillatory zoned plagioclase
crystals surrounded by a glass matrix. A basalt and rhyodacite
sampled from the Galapagos are on loan from the Smithsonian Insti-
tution. A quartz latite sampled from the Superstition-Superior
volcanic area in central Arizona are on loan from J. S. Stuckless

(U.S. Geological Survey, Menlo Park, California).

ANALYTICAL METHODS

Grain Selection

 

Three principle factors were used to select oscillatory
zoned plagioclase crystals for analysis. These factors include
clarity of oscillatory zones at the 2 micron scale, clean and rela-
tively unaltered grain boundaries and a clear glass matrix adjacent
to the plagioclase crystals.

The orientation of the zoned plagioclase crystals within
each rock sample was assumed to be random. Therefore, these samples
were cut at three mutually perpendicular directions to maximize the
probability of cutting plagioclase crystals perpendicular to compo-
sitional zones. Optimal crystal orientation for chemical analysis
of zone patterns was when the concentric zones within the crystal
are perpendicular to the plane of the microscope stage. To deter-
mine if the plane of concentric zoning lies perpendicular to the
plane of the stage, each crystal was checked using the Universal
Stage. Deviations of from 5°-lO° were tolerated and these crystals

were also used for chemical analyses.

Procedure
Quantitative compositional data for this study of oscil-
latory zoned plagioclase was generated using a three spectrometer

ARL EXM microprobe set at an accelerating potential of 15 KV and

11

12

100 nanoamperes beam current. The beam spot size was approximately
.5 microns in diameter. The LiF, RAP and ADP detector crystals
were used to measure the Ka peaks of Ca, Na and Si respectively.
Samples of 100% albite, 95% anorthite and quartz were used as
standards; both peak and background counts were recorded for each
standard.

The three spectrometer microprobe allowed simultaneous
analysis of Ca, Na and Si at each point along the line traverse.
A line traverse is generated by moving the crystal in a series of
equally spaced steps, two microns apart, under a static electron
beam while recording x-ray signals. Several line traverses were
generated such that they began in the adjacent glass and extend
across the crystal-glass boundary into the crystal. Other line
traverses include portions of the glass matrix on either side of
the crystal and extend across the entire crystal. These line
traverses were at right angles to the concentric zoning patterns

in the crystal, i.e., in the direction of concentration gradients.

ANALYSIS OF DATA

Compositional Gradients

 

The first part of the analysis was to establish the
presence or absence of concentration gradients in the glass matrix
immediately adjacent to the oscillatory zoned crystals;

A plot of the line traverse was generated such that the
independent variable, distance, was plotted against counts per
second of the three dependent variables Ca, Na and Si. Inspection
of the glass matrix portion of the plot was then carried out to
determine if the chemical variation of Ca, Na and Si constituted
concentration gradients.

The variation of intensity of counts per second of Ca, Na
and Si at the crystal-glass boundary depend on the bulk chemical
difference between the crystal and glass, the presence of concen-
tration gradients in the glass, width of crystal edge and the type
of crystal-glass boundary. Electron scatter at the crystal-glass
boundary is minimal due to the highly polished thin-section surface.

In Figure 2 the chemical difference at the crystal-glass
boundary is skematically represented for three types of crystal-
glass boundaries. The effective width of the crystal edge depends
somewhat on the slope of the crystal-glass boundary. In the
traverses examined, the slope was steep enough to have negligible

effect at a distance X from the apparent edge.

13

14

The effective width of the crystal edge is defined as the
difference between the apparent and the effective crystal edge.

The apparent edge is the surface intersection of the plagioclase
crystal with the glass matrix. The effective edge is at a distance
X away from the apparent edge. The distance X is dependent on the
type of crystal-glass boundary, the area activated that produces
x-rays under the electron beam and the depth of x-ray penetration.
The effective width of the crystal edge can be measured with a
micrometer under high magnification. The area that produces x-rays
under the electron beam and the depth of x-ray penetration can be
calculated.

The area activated on either side of the spot beam is
approximately 1 micron. The depth of x-ray penetration is dependent
on density and the accelerating potential. Assuming a density of
2.65 gm/cm3 and using an accelerating potential of 15 KV the depth
of x-ray penetration is approximately 2.5 microns.

The effective crystal edge width of Type 1 boundary is
the apparent edge (minimum edge width, Table 1) plus 1 micron.
Although no Type 2 boundaries were observed on the crystals tested,
the effective edge width is the apparent edge width plus 1 micron.
The effective crystal edge width of a Type 3 boundary is the appar-
ent edge plus 2 microns. Since the slope of the crystal edge,
extending under the glass matrix, is steep this width would compen-
sate for any summing of chemical variations of both crystal and

glass at the boundary.

15

A comparison of the maximum edge width limit (maximum limit
of crystal edge extending under glass matrix) and the effective
crystal edge width in Table 1 shows a 1-4 micron difference. Since
the minimum edge width (apparent edge) plus a distance X is suffi-
cient to compensate for any summing of chemical variations of both
crystal and glass, the effective edge width is utilized to define
the boundary between chemical variations due to the crystal and
those due to the glass. Chemical variations within the area between
the apparent edge and the effective edge are due to the summing of
both crystal and glass chemical variations and are referred to as
the "edge effect."

If the chemical variations present at the crystal-glass
boundary are due to concentration gradients in either Ca, Na or Si
then one would expect to find these, similar to those skematically
represented in Figure 2(b), to extend away from the effective edge
into the glass matrix.

For each crystal-glass line traverse the length of the line
traverse in the adjacent glass matrix, type of periphery zoning and
the presence or absence of Ca, Na or Si concentration gradients are
summarized in Table 2. Inspection of the 23 line traverses, shown
in Figures 3-15, reveal that concentration gradients are present
only in the glass matrix adjacent to crystals which exhibit oscil-
latory zoning at their periphery.

Eight of the line traverses are adjacent to the periphery

of oscillatory zoned crystals and six of these show the presence of

16

.mmpcmucaom um cowpmwcm> quwEmcu
can mmaxh zgmucaom mmmpm-_mumxcu we upweocm ovmemxm

.N wczmvu

17

a2

\1IIJ8
III/

 

“cameo

25530

“Boom 3 .95

A6393 fawn coszEcon >3; u a

\ «2
)\a0

 

 

 

 

 

 

 

 

 

Eomca
:0: 2:50:00

N mmDOE

mono 2502.0" 20
mono .ceaanuuaoa

.2353
3290

Earobvoaszoa

 

 

o a
a iiiii 11...!
Eau
new :0
320.2103: O.~

 

63
each :0

Eu

 

.w .P

l‘rl.III-Ivnll

 

can :0

 

“oomuxv
{uncaom .0 on:

18

an Si concentration gradient in the adjacent glass matrix. One of
these six also shows a definite concentration gradient in both Ca
and Na.

Fifteen of the line traverses are adjacent to the periphery
of normally zoned crystals and none of these show concentration

gradients in either Si, Ca or Na in the adjacent glass matrix.

Quartz Latite

 

To summarize, three of the five line traverses adjacent to
the periphery of oscillatory zoned crystals show the presence of
concentration gradients and six other line traverses adjacent to the
periphery of normally zoned crystals show the absence of Si, Ca and
Na concentration gradients in the adjacent glass matrix.

Examination of crystal 1 (Plate 1) reveals six, evenly
spaced, fine oscillatory zones within the perimeter zone of one
crystal face. The perimeter zone of the remaining crystal faces do
not contain these fine oscillatory zones. The preferred location
of oscillatory zoning for one face over other faces may be due to
the chemical environment immediately adjacent to the growing crystal
and the growth kinetics of individual crystal faces.

Two line traverses were generated across the oscillatory
zoned perimeter of crystal l (Figure 3). One line traverse, l-bb',
shows a Si concentration gradient present in the adjacent glass
matrix. The other line traverse, l-aa', shows anomalous behavior
of Si at the crystal-glass boundary and in the glass matrix. Both

l-aa' and l-bb' show no concentration gradients in either Ca or Na

19

in the glass adjacent to the crystal-glass interface. A third line
traverse, l-cc', across the normally zoned perimeter of crystal 1
shows the absence of Ca, Na and Si concentration gradients in the
adjacent glass matrix (Figure 3).

Crystal 2 (plate 1) is normally zoned at its periphery.

Two line traverses 2-dd' and 2-ee' across a crystal-glass boundary
show the absence of Ca, Na and Si concentration gradients in the
glass matrix (Figure 4).

The normally zoned perimeter of crystal 3 (Plate 2) contains
14 oscillatory zones of variable width. These oscillatory zones
are located between the crystal edge and the first abrupt zoning
discontinuity. 0f the 14 zones 5-6 are concentrated at the crystal
perimeter. Line traverse 3-ff' (Figure 5) shows the absence of Ca,
Na and Si concentration gradients in the glass matrix adjacent to
crystal 3.

In crystal 4 (Plate 2) there are 11 oscillatory zones of
variable width located between the crystal perimeter and the first
major zoning discontinuity. Both line traverses 4-gg' and 4-hh'
across the crystal-glass boundary of the same crystal face exhibit
Si concentration gradients in the glass matrix. Line traverse 4-hh'
also exhibits definite Ca and Na concentration gradients in the
.adjacent glass matrix (Figure 6).

In crystal 5 (Plate 2) the normally zoned outer boundary
contains no oscillatory zoning. Line traverse 5-ii' and 5-kk'
across the same crystal face did not show Ca, Na, Si or Al concen-

tration gradients in the glass matrix adjacent to the crystal-glass

aooo.

.
sooo.
(:PSW
20004

J

1000,

elieciwe edge

r"~/"m./\__A/\

\

\

GLASS

 

10 20 so do an ac
DISTANCE (microne)

 

4

 

 

20

 

 

 

 

 

 

 

 

 

l-aa'
effective edge
I'M
40004 p ‘ epperenl edge
I i.
. / i
A Ad 1
soooi‘»1 Jr‘vnr’ 1
i
era) 3
l
v P\ A I
2000‘ \P ‘V J ”vs“
i
1000. GLASS CRYSTAL
L h»- s a" E
o 10 20 so on as an
DISTANCE (microns
1-bb' 1-cc'
400M effective edge
loan"!!! .6“. J \ l‘ ‘ Inherent Odo.
’\ ’-
aoool ‘\\1\P
.4
\ cps. “a,” \
KIM.“ \“ ’
3‘ 2000 I 1 , \.‘~A\-‘
8i
cnvsrAL * GLASS CRYSTAL
sooo.
c. ‘ I.-'~'-‘\\A. c.
”.s.» .o"..° 1’ \
I“ ' ' b “0’ W‘
’ NI 0 ’ N.
70 o ' ' sh ' 5' ea 74* do
'°on§imoe (microge) 7°
Figure 3. Quartz Latite Crystal 1.

 

21

2+dd'

 

 

elleciive edge
8000,
A {4‘
J“ i l\ '
4 V \’J\\' !
'1 000010111 edge
7000. 1
\
‘1
I
l
‘.
.0001 ! ',’~\
I Li".
H
Si
CPSIL
30001
Ne
I
2000. GLASS .
F‘\ l.\,".
‘\ 0' C.
J l \J
l
1000. :
W ‘
. AH CRYSTAL
-.I \-’n‘-‘-’r‘I [

 

1'0 2'0 3'0
DISTANCE

4'0 50 00
(micronel

 

 

- A .‘l
\" "Afl'l \v“
i
7000
may cuss
CPS1

00001

 

i
1

I
l
i

eIIecIive edge

I

I

‘J

 

 

 

2-ee'

appereni edge

.r-w’

, I
t-’\J

«
\VA"! .1".
V

s 1

CRYSTAL

 

 

Do

I

 

 

0 20 a
DISTANCE

Figure 4. Quartz Latite Crystal 2.
3-ff'
8000. eIteche 0090
)MA
v". I/\\
7000, U -\ A
\FH eppeveni edge
1 I
.000, k A
‘\ 5““
CPSJ- A‘\/‘ Di
30001P
1 M
CLASS
2000.
, "~f‘1 I.
‘ i ‘43.
1000. ' '1
r1 " carom.
l 1
.l\’~-’ \—I
o I V v V v v
0 1o 20 so 40 co 00 10
DISTANCE I microne)
Figure 5. Quartz Latite Crystal 3.

40 50
(microns)

22

4-99'
8000. eIIeche edge
A
l \V.’ K f‘o IN
V V '\.f\.‘.“
7000. V\
\d"\V epperentedge
\
1
6000 ‘
1 Ah.‘ I‘M/K.
U
‘L Si
CHI;
4
. ‘f‘\
I‘ \ -
‘ GLASS f ‘\-” \
Ce
2000.
d
CRYSTAL

 

 

 

 

 

2'0 30 4'0 5'0 0'0 To 00 00
DISTANCE (microne)

 

 

 

 

' 4-hh'
eihcuveIKMn
eooq
A
. V'INJ\
\ apparent odge
7000 X '
4 \.A
~\\
0000. ww—
Si
i
C PS " _
l GLASS
Ne
3000,
1
; CRYSTAL
2000
rsl‘VA\\
0 \’~
‘ ‘A 3” Ce
1000 ‘.
I IN! ’
1
I
. I
1". .’
VW“!
o v v w v v v v
0 10 20 30 40 80 00 70

DISTANCE (microne)

Figure 6. Quartz Latite Crystal 4.

23

    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O O .
5-11
30001 eIIeche
edge too-rem edge
CPs‘ ,
I" Iv
’ ‘ 0 \."'~./
2000. I V " AI
cuss :M.
1 A
. I 1
1000‘ AWJT’NP’\.J '\’\‘,/\-.‘
Ce
‘ /-¢\\ Iav"\’\o-
, _ CRYSTAL
0 V2 _ L a -
0 10 20 00 40_ 00 00 70
DISTANCE (rmcrons)
O O.
- i
5 JJ 5-kk
' eHecIive edge 1
0000. 0000
\ F. 1
u; eiieciive edge
‘ l ,5. I
‘ Id ‘1 J"- 0000reni edge
‘I s ‘
e. ‘ ’
. - 1
i ! ‘. V\,’\/‘N\ epperenI edge
"\ r’hhr‘x/J 1 l.
|
”“1 1.»; 01 0000. ‘.\
A A“
.1 l I
cps .r crsL ‘J V 01
3000, M moor cuss
cuss M
Ne
1
) CRYSTAL
2000) f‘ 2000.
0
' «r .
1 .' V\"'\ ‘ \\’A"’\’\
i ‘1' \
l C Ce
1000 I" : ' 1000.
’sJ ‘\’ .\\ " I..l I
. V CRYSTAL {
\V" ‘\«-“’I
O 1 , - fl r a a _ ' v
0 10 20 00 40 00 00 70 o 10 20 0'0 4'0 0'0 0‘0 7‘0
DISTANCE (microns) DISTANCE (microns)

Figure 7.

Quartz

Latite Crystal 5.

24

interface. A third line traverse, S-jj', across the crystal-glass
boundary of another face of crystal 5 did not show the presence of
Ca, Na and Si concentration gradients in the adjacent glass matrix

(Figure 7).

Rhyodacite

 

To summarize, the line traverse adjacent to the periphery of
oscillatory zoned crystals show only the presence of Si concentration
gradients in the adjacent glass matrix. The remaining line traverses,
across normally zoned peripheral boundaries, do not exhibit Ca, Na or
Si concentration gradients in the adjacent glass matrix.

Line traverses across the normally zoned periphery of crystals
l, 2 and 3 (Plate 3) show the absence of Ca, Na and Si concentration
gradients in the glass matrix adjacent to the crystal-glass interface
(Figures 8-10).

In crystal 4 (Plate 3) there are 9-10 oscillatory zones of
variable width located between the crystal perimeter and the first
major zoning discontinuity on one crystal face. Line traverse 4-ff'
across the crystal-glass interface of this crystal face exhibit an
Si concentration gradient but no Ca or Na concentration gradients
in the adjacent glass matrix (Figure 11).

In crystal 5 (Plate 4) there are 4 faint oscillatory zones
located between the crystal edge and the first major zoning dis-
continuity and extending around the circumference of the crystal.

Line traverse S-gg' across a crystal-glass boundary exhibits an

Si concentration gradient in the adjacent glass matrix (Figure 12),

25

but no Ca or Na concentration gradients were observed in the glass
matrix.

Crystal 6 (Plate 4) is normally zoned at its periphery.
Line traverses 6-hh' and 6-ii' across the crystal-glass interface,
of two different crystal faces, did not show Ca, Na or Si concen-
tration gradients in the adjacent glass matrix (Figure 13).

In crystal 7 (Plate 4) there are 12 oscillatory zones of
variable width, located between the crystal edge and the first
abrupt zoning discontinuity, which extend around the entire circum-
ference of the crystal. Line traverse 7-jj' across a crystal-glass
boundary reveals the presence of an Si concentration gradient and
the absence of both Ca and Na concentration gradients in the adja-
cent glass matrix (Figure 14).

Two line traverses, 8-kk' and 8-ll', across a crystal-glass
interface of the normally zoned perimeter of crystal 8 (Plate 4)
showed the absence of Ca, Na and Si concentration gradients in the

adjacent glass matrix (Figure 15).

132511.:
The glass matrix of the basalt sample was devitrified and

it was not possible to determine if concentration gradients were

present in the glass matrix adjacent to the zoned plagioclase

crystals.

 

26

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 i
l-aa "mm. l-bb
I .dwg’l«l’\\
\J \.
eIiecIive edge 1 00001001
4000. 4000. i 009.
epoereni edge ‘A
\ /. I"~ ‘\ .’
I .A \vN/\\,-I W/\ « \‘J’\! .I \
1 ‘1: 81
3000 1".\’TN\’~- 0000. I ’-
V Si L!
0981 crs(
cuss L CRYSTAL ."
1
_ I
2"” cuss CRYSTAL 2°°°‘ ‘ ' ‘v’\ 'z’ c.
I "
1 4 ' \\I
I
l'A"""“"‘ .-0."\ ,~’
1000. ”0 c: 1000. 1 - /\
I Ne
I‘ if
1-,~/” “’“”J JKJ'””"‘-v~ l~uNCArxj
N0
0 a g. - i - - o - f - - f
0 1o 20 30 40 so 00 70 00 0 10 20 30 so so
DISTANCE (microns) DISTANCEUMcrono)
F1gure 8. Rhyodac1te Crystal 1.
2-dd'
CHOCIW. edge
0000.
h
' ‘I/ M’\’
Z'CC ’ \' eppereni
4000. 4000. 13 “'9'
01100111100000 coo-rented” 'I ’
‘ CPS 1 ‘ 5M,—
“Avf‘ ‘ ISi
0000. H‘ 3000‘
VV¢KV'\"’\’
CPS‘
SI
2000‘ 2000. cuss CRYSTAL
cuss CRYSTAL .
A o‘ ’
I P-—’\ J- I \ I V-
1000 “’\’ ‘ vc. 1000. '4 \l C.
I. ‘ ’A'“‘I"-v" F,“ W
‘\’K\.o”-‘l"' U "M i/ N.
W/ N.
° ' ' 0° ' 2' 0'0 4'0 0'0 0'0 7‘0
0 1'0 2'0 00 7 0'0 00 70 10 0
DISTANICE 8micr0ns) DISTANCE (microns)

Figure 9.

Rhyodacite Crystal

2.

27

 

 

 

1
3-ee
5000 effective
‘ edge
!\. :"r-II
A") Aw. I
\.I \_/ - " eooereni
4000. f 1 edge
I
e “\“‘../¢II\;\
CPS 1 Si
0000.
GLASS ‘\
(V‘V’ "’\
I Ce
2000. 1
l
I ,' CRYSTAL
\\4”\fl /‘\’\A\’J
1000 ‘W
W
Wk/ N.
O - . r . .
0 10 00 so so 00 70

 

270
DI STA NCE (microns)

Figure 10. Rhyodacite Crystal 3.

4-ff'

 

 

 

 

 

 

 

 

eiieciive edge
4
M ’ I 1
0000. A! \M’ \~\ 5'99
V‘.
his
‘. epoereni l eiieciive edge
1 edge
4000 k. ‘000.
1 \’\~...‘
81 eppereni
CPS1 ‘ ,’\.4/\ .6
‘0
VJ» .
0000. 0000. \‘
A A Krfirr‘vfim
I (IL 00 cevsr L CPS1 01
2000. 2000.
,q ' cuss cnvsm.
\c"\l~t~ ‘
1
.I‘ 1“.“ ‘\AV\ ""‘4 1' Ce J %/\..’I‘\... _“
1000. 1000. ' c-
1
W M
1 j a. 1 l
W N.
Vn‘ .I -A‘A
0 fi . - - . .
0 70 2'0 0‘0 1'0 0'0 00 7'0 '10 2'6 00 10 00 00 70
DISTANCE (microns) DISTANCE (microns)

Figure 11. Rhyodacite Crystal 4. Figure 12. Rhyodacite Crystal 5.

 

 

28

 

 

6-hh'
4
room
eIIeche edge
-—~I’\ "
6000. ”T \’\.’\’ \ gpperent edge
A
1
CPS ‘ I
|
NVI.\ A
0000. LP " '
L 3'
1-
i C'-
OLA S S ‘,\""\’_- ’\II
1000‘
Ne
‘ I CRYSTAL
‘ . “\a'--
o ' fl .
0 1o 00 00 70

2'0 50 4'0
DISTANCE (microns)

Figure 13.

0000. m EV
'? ‘x
A ‘11! \VN‘

0000) N!
CPS4
0000
ii GLASS
11'
1000

Figure 14.

1 011001110 edge

 

 

 

 

 

 

 

C U '
6-11
70002
eIIeCIwe edge
0000. V'" !\’\__.’
V'W \1 eooerent
I 00
cps. 3 V
'I
0000. I A 1‘. ,-
‘L V’Ky," v ‘31
1P
CLASS \,.I’\ Ad!
10001 NI
. CRYSTAL
I I\ : :~
I
0 -
0 70

eopereni edge

IOCV

l

CRYSTAL

1'0 2'0 0'0 1'0 0'0 0'0
DISTANCE (microns)

Rhyodacite Crystal 6.

,(\. ’e

M

~I\’Iv\[ \l’ '

 

Ce
4 W:
‘sd"\."‘\".o~fl
o V V v w fi fi fi
0 10 20 30 40 50 00 70 00
DISTANCE (microns)

Rhyodacite Crystal 7.

29

8-kk'

7 0 O 0‘ eIiecI ive edge

1 . I‘-\ -
\o [\J v \’\.r'\“

 

 

 

 

0000. 1
I
eooereni edge
01201 '1
)- M I! f’
00001 “V v‘\~J L’
‘. GLASS 3'
1 N0
1000 M
‘ /" I Ce
I
m,/
.,1 CRYSTAL
’sa‘fv'\l\ld
o . a , r - - . .
0 TO 20 30 40 50 00 70 SO
DISTANCE (microns)
8-11'
4
“10011110 edge
7000.

. AIR/\VP‘VK

C

O

O

.0
’0’”

eopereni edge
(A
\Ao".
cuss 3'
Ce
A‘,v\/‘"
1o 0 0.

Ms
M

d a
NVN'vos/‘N"’ CRYSTAL

3L

 

 

 

 

1‘0 2'0 0'0 70 0'0 0'0 1'0 0'0 1? 0
DISTANCE (microns)

Figure 15. Rhyodacite Crystal 8.

30

.mmnxh xgaucaom Pmumxgoummm_o new cuuwz FmHmXLu m>wuuwwdm

.F mpnmh

31

QGQCDVCOVO

oaaoaaov-pconv—oco
ESZoaoo«ee<Dv'¢v<D
ouwevwnwcvvcvmcvv

Deco
v-v-

wh.0<OO>Im

NP xxum
\
o, \:.m
o. L.-.“
\cs..v
\ua..v
\d...m
\mouu
\uuuw
‘00:.
won.
Wan.

VP
or

”Contflflv-wv-GGDM

vqcocpecomcofi
v<n¢>9~t¢~38§3§

chv<¢wcvucu

Eav
AxaE .zEV
mth<4 NbI<DC

 

wa>h Ibo-3 woow wh_2.4 :22;
>m<DZDOm 4<hm>¢0 w>.._.ow..._u_w woom 4<hw>m0 wmmw><¢h NZ...

32

.mucmwcmgw cowumgucmocou xwgpmz mmmpo new maah mcwcoN xgmsawgma

.N mFQmH

33

“.3560 an; 62:53; ..

 

z z 2 EB: we \ :..m
z z 2 .52. 3 wt;
2 z > 23 «v . 2-5
z z z .59. on . .To
z z 2 ES: «v $.50
z z > 93 «c 03$
2 z > 36 av .:.v
z z 2 EB: ov ‘00.,”
z z z 6.0: on \uu-~
z z z .52. on be.“
2 z 2 E5: on be;
2 Z z :20: E: @v .aac—
th<00>Im
z z z :20: ov 2.3.6
2 z 2 ES: on \ .:-m
z z :2 ES: 3 ‘ z -m
> > > omo N1 2sz
z z > one. No .3-..
z z z omo mm L70
2 z z 8.0: av .00..“
z z 2 .52. «v .20-...
z z 2 EB: om .ooL
z z > 9.3 cc an;
2 z 2 80 E: me ha.—
m2 m0 .m mat. Eco mam-o :5 3.72.. th<30
Azx>v bzw.o<m0 02.20N wwmw><mh

20....(Eh2w0200 >men:mwm wz... m0 Ihmuzwg wmmm><mh m2:

34

Pattern of Oscillatory Zones

In the second part of the analysis the oscillatory zone
patterns of each plagioclase crystal were examined to determine if
the patterns between crystals could be correlated.

First each line traverse extending across a zoned crystal
was subjected to the FILTER program (Appendix). The FILTER pro—
gram prepared the data for further analyses by performing three
functions.

First, it deleted the segments of the line traverse extend-
ing beyond the crystal-glass boundary into the adjacent glass
matrix, leaving intact the portion of the line traverse represent-
ing the plagioclase compositional stratigraphy.

Second, it searched for data points throughout the remain-
ing portion of the line traverse which did not exhibit mutual var-
iation in all three major elements. These points were assumed to
represent crystal imperfections or impurities on the crystal sur-
face. Because Cross-Correlation and possibly Fourier analysis were
contemplated as analytical tools (both require data points at
equally spaced intervals) these points were not deleted from the
line traverse but were instead displaced to a new position. The
new position was produced by averaging the two data points on
either side of the displaced point. The percent of these points
per line traverse varied from 0-3% which is considered insignifi-
cant.

Third, the FILTER program transformed the line traverse

from counts per second to weight percent anorthite using methods

35

and correction factors of Bence and Albee (l969). This data was
then presented in the form of an orthogonal plot of weight percent
anorthite verse distance (in microns) across the crystal (Figures
16-29).

This data (in the form of a plot) was evaluated to deter-
mine if the zone patterns between crystals could be correlated.
Correlation of the compositional stratigraphy between crystals is
based on a match of both the number of zones and the zone width.
However, before correlation between crystals could proceed a good
correlation must exist between optically observed zone patterns and
those exhibited by the line traverse. This is necessary because
recording x-ray signals at two micron increments may not adequately

define zones which are from l-4 microns in width.

Quartz Latite

 

In the quartz latite 2 of the 4 line traverses show limited
agreement between optically observed and line traverse zone patterns.
In both plagioclase crystals, line traverses 4-ab and 5-lm, there is
a distinct increase in weight percent anorthite in the portion of
the line traverse representing the core of the crystal. Within the
core of each crystal there were 3-4 large zones varying in width
from 50-75 microns. The cores and the 3-4 large zones within the
cores were the only zone patterns which could be correlated between
line traverses 4-ab and 5-lm (Figures l7 and IS). No further cor-
relation of zone patterns between these two line traverses were

found.

36

The zone patterns represented by the remaining two line
traverses 2-ab and 6-ab (Figures l6 and 19) could not be correlated
with respect to the number of zones and zone width observed opti-
cally.

Although there was a lack of agreement between the optically
observed and line traverse zone patterns, optical examination of the
zone patterns in these four crystals (Plates 1 and 2) and examina-
tion of the line traverses reveal a large variability of both zone
width and the number of zones per crystal. Therefore, except for
the crystal cores of line traverses 4-ab and 5-lm, there appears to

be poor correlation of zone patterns between these crystals.

Rhyodacite and Basalt

 

In the rhyodacite and basalt all the crystals examined show
a lack of agreement between their optically observed zone patterns
and their respective line traverse zone patterns (Figures 20-29).

The lack of agreement may be due partially to the small
chemical variation within these microphenocrysts. In most cases
the chemical variation across the crystal is less than l0 percent
anorthite. Although a small change in weight percent anorthite
between zones may be observed optically, this change may not be
recognized as the termination of one zone and the initiation of
another zone in the line traverse.

As stated earlier, recording x-ray signals at two micron

increments may not adequately define zones which are from l-4

37

microns in width and this may contribute to the lack of agreement
between optically observed and line traverse zone patterns.

As displayed in Plates 3-5 there is a wide variability in
the number of zones per crystal and the zone width of the crystals
sampled from the rhyodacite and the basalt. Again by inspection
of the optical zone patterns it was concluded that there was poor

zone pattern correlation between crystals of each rock type.

38

cm;

3:32;: woz<hw
2.5

IF h

 

5
com

elm 2:3 2:30

or wmawl
09 o

h -

 

ON

om

39

no.1 22a.— 23:0

$53.5 moz<hm .o C manor.
3.8 8.9 8.9 8.: 8w; 8.9. can one 2.: co... o

P

on

 

 

I

on

 

 

 

NV‘ilM

40

81m maze; utaao

3:90.2v wow—(baa mp mane:
coo , coo r omv . cow . 0
ON

 

cow. . . cob. . .

.on

 

 

.0.»

on

 

NV%1M

4l

 

 

 

 

no.0 32.: 2330

3:92.»; moz<._.m_¢ aw manor.

one can on: com com 8. o
op
ram
do

 

egg???

dv

 

NV'ilM

80,

70‘)

WT% AN

604

 

50

42

 

0 85'0
FIGURE 26" T

. 1bo
ANC‘E (microns)

Rhyodacite 2-pq

90,

80.

WT%AN

 

70

 

0 2'5
FIGURE 21-
Rhyodacite 3-ab

5'0
DISTANCE (microns)

7'5-

Too

43

80.

WWI/W

 

 

 

 

 

 

 

 

 

 

Z
<
““75.
.—
3
7o . .
° 25 50 . is 100
FIGURE 22 DISTANCE (micronS)
Rhyodacite 4-tu
80.
Z
“ \
aQ'IS‘J
.-
3
70 ' '
O 25DISTAIvCSEOI ‘ ro ‘s)f5 10°
FIGURE 23 ' "“c "‘

Rhyodacite 4-ut

85d

 

44

 

 

 

80 .
z -I/
< / //\j\/Av
59
I'-
E
75,
7O 1 1
o '5 NCE ( . ) 75 760
mICFOI‘lS
FIGURE 24 0's“)
Rhyodacite 4-sr ,
90,
Z
<
a‘?
)-
2
so,
70 '
0 so , 100 150
DISTANCE (mIcrons)

FIGURE 25
Rhyodacite 8.cg

45

 

 

 

 

 

 

 

 

 

BO.
2
<
3275
I'-
a l
70 .
0 50 - 60 , 150 560
Base” 1-xo
80..
Z
<
327
HQ.
3
60 t .
0 5ODISTI\NC13(E)O( ' r nsjéo 200
I‘m
FIGURE 27 C 0

BasaltI-pz

46

8 Q)

WT%AN

W

 

7O

 

o 5‘00.STAN IDO( .. 1'53 200
mICTOl'I
FIGURE 28 CE

Basalt 2~kj

90

80.

WT%AN

 

 

7O

 

° 1 DISTcAONCE ( r- 26°
FIGURE 29 mleQflS)

Basalt 3-gb

DISCUSSION

The presence of concentration gradients in the glass matrix
adjacent to oscillatory zoned crystal faces is predicted if oscil-
latory zoning in plagioclase results from constitutional supercooling.
This data provides support for the constitutional supercooling model
as a mechanism by which oscillatory zoning is produced.

Evaluation of the data presented in the Compositional Grad-
ient section reveals that concentration gradients are present in the
glass matrix adjacent to crystals which exhibit oscillatory zoning
at their peripheral boundary. The peripheral boundary is defined as
that area between the crystal edge and the first abrupt zoning dis-
continuity. In all line traverses exhibiting Si concentration grad-
ients only one, quartz latite 4-hh', also exhibits a definite Ca and
Na concentration gradient in the glass matrix. The absence of Ca
and Na concentration gradients at crystal-glass boundaries, where
Si concentration gradients are present in the glass matrix, may be
due to the higher ionic mobility of Ca and Na compared to Si and Al.

In both the quartz latite and rhyodacite the distribution
and number of oscillatory zones within the crystal's peripheral
boundary varies significantly from crystal to crystal. The number

of oscillatory zones contained in the peripheral boundary varies

47

48

' from 6-l4 for plagioclase sampled from quartz latite and from 4-2
for plagioclase sampled from the rhyodacite.

In all three rock types oscillatory zoned plagioclase
crystals occur along with plagioclase crystals which are not oscil-
latory zoned. If the oscillatory zoning were produced by general
environmental variables, all of the plagioclases would contain
oscillatory zones (unless they were xenocrysts).

A third supportive point is that there are no concentration
gradients observed in the glass matrix adjacent to normally zoned
crystal perimeters, even when these crystals were in the neighbor-
hood of other crystals exhibiting oscillatory zoned perimeters.
This suggests that a favorable chemical environment necessary to
produce oscillatory zoning in the former did not exist in the melt
adjacent to those growing crystals.

A fourth point of prime importance is that at the boundary
of two crystals, quartz latite 1 (Plate 1) and rhyodacite 4 (Plate
3), oscillatory zoning is isolated to just one crystal face while
the remaining perimeter of each crystal is normally zoned. This
suggests that the growth kinetics of individual crystal faces and
the boundary layer melt are the controlling parameters which deter-
mine if oscillatory zoning is produced in a crystal.

The oscillatory zoned crystal faces may represent repeti-
tive advance and completion of the planar growth interface, here
the growth rate of the crystal exceeds the diffusion rate of the
solute in the melt. The planar nature of the interface may inhibit

nucleation on those crystal faces exhibiting normal zoning and

49

solute diffusion rates in the melt exceed crystal growth rates,
which is indicated by the absence of a concentration gradient in
the adjacent glass matrix.

Finally, in all crystals except two, quartz latite 4-ab
and quartz latite 5-lm, there is no zone pattern correlation between
crystals sampled from the same rock type and in many cases there is
a variation in zone pattern in different crystallographic planes.
It is also interesting to note that the average composition of these
neighboring zones is relatively constant. This may indicate that
crystallization occurred under nearly isothermal conditions.

Although there was not a good correlation between optically
observed zone patterns and those exhibited by the line traverse,
the lack of agreement is not thought by the author to be due to
analytical error for two reasons. First, line traverses across the
larger phenocrysts, sampled from the quartz latite, do show agree-
ment between the optically observed and line traverse zone patterns.
This agreement is found at the abrupt change in weight percent
anorthite at the core and not at other points of more subtle change
in composition. Second, parallel rhyodacite line traverses 4—tu
and 4-ut (Figures 22 and 23) across a single crystal and located
along an identical path exhibit good agreement with respect to the
recorded chemical variation when the plots are superimposed on each
other. This demonstrates the repeatability of the method.

The author's interpretation of this data presented above
is that in the quartz latite there is limited zone patterns cor-

relation between two crystals (4-ab and 5-lm) and no zone pattern

50

correlation between crystals sampled from either the rhyodacite or
basalt.

This data is interpreted to mean that the parameters which
control oscillatory zoning are localized about the crystal-melt
interface and are not a function of changes in general environmental
variables in the magma chamber.

The presence of concentration gradients in the glass matrix
adjacent only to the periphery of oscillatory zoned crystals provides
the strongest support for constitutional supercooling as a mechanism

for oscillatory zoning in plagioclase.

CONCLUSION

In conclusion, this study provides four major lines of evi-
dence which support the constitutional supercooling model as a
mechanism by which oscillatory zoning in plagioclase can be produced.

The first is the presence of concentration gradients in the
glass matrix adjacent to only those crystals which exhibit oscil-
latory zoning at their perimeter. Second, no concentration gradients
are present adjacent to normally zoned crystal faces. Third, at the
boundary of two crystals, quartz latite l and rhyodacite 4, oscil-
latory zoning is isolated to just one crystal face and the remaining
perimeter is normally zoned. Fourth, there appeared to be only
limited zone pattern correlation between crystals sampled from the
same rock type.

Therefore, this study provides evidence which is interpreted
to support constitutional supercooling as a viable mechanism for

oscillatory zoning in plagioclase.

51

APPENDIX

52

900
899

$01

902

1237

2000
11

2001

1000

FRCGRAP FILTER(INFur,oUTFur=65,TAFE2,TAFE7,TAFE1c,TAFE11)
REAL NICRCN,NA
INTEGER 21c1~c
REAL MAX,‘IN
REAL u1crcr,nxcacr
REAL Locrxc
INTEGEF TOKFK,NPROC,TOF,90T
INTEGER PFN
cowrsx/ErEcrsc/RNF
cc«rtA/EMULATIIEAuo,xs15E,vsxzE,vKIN,er~

'{OnfiON DU!,K1CRON(1002),Sl(1¢02),€l(1002),NA(1502)

CCVVCNIFLOTE/PLOTBF(10‘9)
COFECNITDPIFDE(3C),FFN(6)
laAuosizoo

XSISE=25.o

YSIZE=ZS.O

XFIN:0.0

vr1x=3.c

REIIND 2

CALL PETUPMT<SLTAt67l

REHIND 7

PRINT 900

FURFAT(* ENTER SANFLE°*)

READ 899,?FN(Z)

FCR”AT(A10)

PRINT 901

FOR:AT(* ENTER anls-a)

READ 899,Fru(3)

rPlhT 902

FORMAT(* ENTER TRACE-*)

READ 899,PFh(6)
PFN(1)810HVIV-D&TA

(ILL FFFCPISLTAFE7,FFN,FDF,20)
CALL FFFFP(FD§,“PP“,300)
MAX=0.0

th=1COOGOO0.0

PRINT 1237

FC?MAT(*OREADIKG THE DATA....*)
NIN=D

CONTINUE

NIN=NIN01
READ(2,1000)N1CRDN(N1N),SI(N1N),CA(NIN),NA(NIN)
1F(LIN.E0.1)GOTO 1
15(EOF(2).NE.0.0)GOTO 11
1F(SI(VIN).GT.M!X)VAX=SI(NIN)
IF(CAININ).GT.HAA)MAX=CA(N1N)
If(N&(h1k).GT.HAX)HAX=NA(NIN)
1F(SI(NIN).LT.HIH)MIN=SI(NIN)
If(CF(NIK).LT.MIN)FIN=CG(NIN)
IF(NAINIK).LT.NIN)$IN=NA(NIN)
1F(NA(NIN).LF.0.0)PFILT t,NA(NIN),NIN
IIININ.LT.1000)GOTD 1

FAINT 2000

FOPMAT(* THC Fucv rtTA....*)
CONTINUE

NIN=NIN~1

NIH=(NINIZ)*2

rqqu 2001,v1w

FORKAT(* *,IS,t CAsss READ Isa)
IF(N1N.LT.5)STDF "Too LITTLE DATA"
quMAT<r4.0,3rs.0)

CONSTY=10.0I(MAx-M1N)
CONSTX8(FLOAT(NIN)IZS.O)

53

54

IF(CONSTX.GE.40.C)CONST¥=£0.0
CCNSTX=CDNSTXIHICRON(NIN)

END OF THE DATA INPUTTING PORTION...

”00

PRINT 3901
3901 FOP“AT(¢003 vou CANT TO SEE THE UNFILTERED FLOT?*)
CALL VETO(TOKER,NFPOC,XNUM)
IFCTOYEN.E0.1FN)GOTG 300
CALL PLDTCP(N1N,PAX,HIN,3,NIN-2,CONSTX,CDNSTY)
300 CSATINUE

C NON HE CONFUTE THE GLASS EDCES....

PRINT 3000
3000 FOFFAT(*CCOVFUTIAG THE CLASS EDGES...*)
BOT=1$TOF=NIN
l1CINC=INT(PICRO~(5)-MICRON(A))
IFCKICINC.FG.0)FICINC=1
MICINC=?IFICINC
IFIFICINC.LE.O)FICINC=1
N =NIN-5
DC 2 LUP=¢,AN
UFUszACNUF+M1CTNC)¢HA(RUFINICI~C+1)
UrDN=NICAUF-NICINC)+RA(Nut-NICINC-T) ~
If((EDT.EG.1).A~D.((UFDN*S.O).LT.UPUF))EOT=NUP
IF((UFUPtS.0).LT.(UFDN))TOP=NUF
CONTINUE

“N

FRINT 3001,RICPONCEGT),MICRON(TOP)
3001 FORNAT( * FOTTOH OF GLASS SEE”S TO BE * ,F6.1, * MICRONS * I
+ * TOP OF GLASS SEEMS TO BE * ,F6.1, * RICRONS * )

PRINT 3002 .

3002 FDRMAT(* DO YOU HISP TO CHANGE THE GLASS LOCATIONS-YIN*)
CALL VETOCTOKEN,NPAOC,XNUH)
IFCTOKEN.E0.1HN)GDTO 20

36 CONTINUE '
PRINT 3003

3003 FORMAT(* ENTER BOTTOM OF GLAss-t)
READ *,NIC80T
PRINT-300A

3004 FDRRAT(¢ ENTER TOP OF GLASS~*)
READ e,MICTOF
IF(MICEOT.GE.MICTOP)GOT0 36
BDT=TOF=1
on 21 N=1,NIN
IF(FICROV(N).LT.MICTOF)TOP=N
IT(MICRDV(N).LT."ICBDT)ECT=N

21 CONTINUE
20 CONTINUE
TDF=T0F-4
C
BOT=80T+L

C THE LOCATION FAS BEEN SET FOR THE TOP AND PCT OF THE GLASS
C
LOCRIC=10000000.0
"TIPST=O
255 CONTINUE
PRINT 2367
2367 TORNAT(*OEEGINNING THE FILTERING PORTION...*)

31

4000

55

N=EOT+6
AVSI=SI(N)
AVCA=CA(N)
AVLA=NA(N)
T0rEN=1HN
NLIST=O
IF (MFIRST.NE.O)GOTO 2926
CONTINUE
wsN+1
AVSIsIAVSI*S.o+SI(~-1)-sI(N-6))IS
AVCAzCAVCAtS.o+CA(N-1)-CA(N-6))IS
Ava-(AVNA¢S.6+AA(~-i)-NA(~—6))JS
DITFSI=ABSCSI(N)-AVSI)*4.0
DIFFCA=IAVCA~CA(~))*1.L

01FFNA=CAVNA-NA(N))*1.A
IF(01FFSI.GT.AVSI)GCTO 31
IFICIFFCA.GT.AVCA)60TD 31
IF(DIFFNA.GT.AVHA)COTO 31
IF(LOC%IC.LE.HICRFN(N))CDTO 31
GOTC 3o
CCLTJIUE
IFITOKEN.E0.1IC)GOTC 32
PRINT Looo,rTC°0v(N),ST(N),CA(M),NA(N)
FOFNAT(* ’,F6.0,* n1c..s:=t,rs.3,,« (Act,re.3,w NA=*,F8.3,* DO YOU

+ LISH TO DELFTE*)

If(TGKEN.E0.1HL.AND.NLIST.LE.NPROC)GOTO 32
CALL VETOITOKLN,NFROC,xNUR)

NLIST=O

.0
.0
no

2924 CONTINUE

IFIMFIRST.ED.O)FOT0 3216
FFIvT t,"MICRDN-”
FEAD *,LOCHIC

3216 CONTINUE

32

IF(TOAEN.EG.TH$)N=TOP
IT(TOKEN.EG.1HS)GOTO 3
IFCTOAEN.EG.1HN)GOTG 3
CONTIKUE

NLIST=\LIST*T

o
9

C HERE HE PLOT THE LINE THAT HARKS THE BAD FOINTS

C

30

3904

C NOW

9876
8765
3902

SI(N)=FVSI

CAIN)=AVCA

NACN)=AVNA

CONTINUE

I‘(N.LT.(TOP))GCTO 3
IFCMFIRST.EG.1)GOTO 3904
HFJPST=1

PRINT *,"bO YOU WANT TO DELETE "ORE POINTS?“
CALL VETOCTOKEN,NPROC,¥NUH)
IFCTCKEN.E$.THT)GOTO 254
CCNTINUE

AF PLOT THE FILTERED DATA IF REQUESTED
PRINT 3902
DD 9876 K=EOT,TOF

IFCNA(K).LE.0.0)FRINT *,K,NA(K), " NA LE 0.0”
URTTECTI,8?DS)r,HICRou(K),SICK),CA(K),NA(K)
FORNATCTX,IS,5FTO.?)
FDRMAT(* DO YOU WANT TO SEE THE FILTEFED PLOT2*)
CALL VETCITOKFN,NPROC,XNUH)
IFCTOKEN.E0.1HH)GOTO 3999

C

C
3999
C

56

CALL PLCTCF(NIN,MAX,MIN,BOT,TOP,CONSTX,CONSTY)

CONTINUE

C NON HE 00 THE FLAGIDCLASE ANALYSIS
C

5000

C

PRINT 5000

FOFFAT(I Do YOU-CANT A PLAGIOCLASE ANALYSIS DONE-I
CALL VETCITOKE~,NFPOC,XNUM)

IFCTOKEN.EO.1HN)GOTO 7000

CALL FLAGCTOF,POT,CONSTX)

C NOR HE STORE THE DITA......

C

6000

7000

1000

2000

1001

C

PFINT 5000
TOFFAT(* DO YOU KANT THE DATA STORED?*)
CALL VETO(TOKE~,AFROC,XNUH)
IFITUKEN.E0.1FY)CALL STORE
CONTINUE
ENO- »
SUBRCUTINE VETO(TOKEN,NPROC,XNUM)
INTEGER TOKEN,NPPOC
PRINT 1000
TORMAT(* ?t)
CCNTINUE
CALL NOELANK
AEAD ZOOO,TOKEN,XNUF,HPPOC
FORMAT(2A1,110)

If(TCKEN.E0.1HV.0P.TOKEN.E0.THN.OP.TOKEN.E0.1HC)QETURN

IF(T3KEN.EO.1HL.OP.TOKEN.EG.1HS)RETURN
FRINT 1001

iORhATC' Y 00 N9.)

ECTO 1

RETURN

EHO

SUFRCUTIHE PLOTCP(NIN,HAX,FIN,BOT,TOF,CONSTX,CONSTY)

INTEGER FIN
REAL HICRON,NA
PEAL 'AX,VIN
INTEGER ssr,ror
crmrow ~,v1cnowrionz),sx(1002),CAIIooz),~A(1002)
CCHMONIFDE/FDP(30),PTN(4)
connovleLoTEIFLDrar(1049)
CALL PLOTSCFLOTBF,1027,0)
CTNSTT=CORSTV*.565
CALL FLI:11I45.0)

C FIRST HE ORAN THE LEGEND

C

CALL NEHFENC3)

C‘LL SY!POL(OQO'°eO'e1h’1OH OONA'OOO"°)
CALL NEUPENCZ)
CALL SYMFOL(0.0,.25,.14,10H ..C!,0.0,10)

CLLL NEUFEN(1)

C9LL SYMBOL(0.0,.5,o14,10HLEGFNO.oSI,0.0,TO)
CALL srseoLto.o,1.o,.28,FrN(£),o.o,10)

CALL svnanLro.o,1.4,.28,Prv(3),o.o,10)

CALL SYWBUL(9.0,1.8,.28,PTN(2),0.0,10)

CALL PLOT‘2e0'1e°p-3) '

C

20

nnfi

nnn

40

50

60

NC

57

N LE FLCT THE AXISES

X-AXIS...

THE

FUT

NCL

CALL PLOT(-.3,0.0,3)

X=0.0

CINTINUE

X=Xt200.0

XCCORDSCGVSTX*(X)

CALL PLOT(XCOORD,0.0,2)

CALL PLOT(XCOORD,'O.2.1)

CALL NUHPEF(FCOGRD’.2,-.£,.44,X,0.0,0)
CALL PLOTCXCOOGD,0.0,3)
IT(X.LT.MICRON(NIN))GOTO TO

Y-AXISCIO

CALL PLDTID.o,-.3,3)

v=0.0

CCATINUE

v=VIzoo.o

YCCORO=CUNSTY*(Y)

CALL FLDT(o.c,VCOCnD,2)

C:LL FLCT(-.Z,YC“OFD,Z)

CPLL RU!DEF(-.a,vCDURD-.2,.1L,vr10.o,9o.o,1)
CALL PLOT(0.0,YCDDRD,3)
IFCY.LT.(HAx-99.0))GOTO 20

THE TITLES IN

CALL SYHBOL('T.O,3.0,.28,3HCPS,90.0,3)
CALL SY"LOL(2.0,-1.0,.Z8,10HSIICAINA ,O.D,TO)

RE PLOT THE LINES...

1:3

DD LO K=BCT,TOP
xccocoscousrroMICFoNIK)
VCCOED=COASTY¢CSI(K))

CALL PLOT(XCOOPD,YCOOR0,I)
1:2

CONTINUE

1:3

CALL NEVPENCZ)

DD so r=sor,ror
XCOOE0=CDNSTX*FICRON(V)
VCCOR0:CONSTY*(CA(K))

CALL FLDTCxcnooD,vcoqu,I)
I=2

CDNTIAUE

CILL CEernrs)

I=3

Do 60 K=Ecr,ror
xccoQD=Cnasrth1csouIKI
vcrcrn=consrv:rv:(r))

CALL PLDTIxcnorn,vcoan,I)
I=2

CONTINUE

CALL FLDT(0.0,0.0,999)
RETUPN

END

996
998
997

87

901
30

31

32
902

50

111

58

SUERLUTINE PLAG(TOP,BOT,CDNSTX)
Inrscsn 109,531
INTEGER EL
DIRENSION IREDCZ)
LOGICAL ITEST
CLF'ONIFLUTBF/PLOTRUF(1DA9)
CCMHON/IN,DYN1(1002),DYN2(1OOZ),DYN3(1002),DYRA(1DDZ)
DIEENSIDN IDCI)
DIRENSION BETA(S),CFS(3).ISTDC3),EL(3),CPSSTD(3,3),H(3),C(3)
IhTEGER LFORPF
DATA ISTDI”AN","AB",”0R"I
DATA BETA/1.11R,1.289,1.162,O.O,O.OI
DATA ICHECKI”Y"I
D‘TA ELI"CA"’”R~‘”'”K .0,
DATA IRED/10thARGIN,13,1L6/
IF(EUT.LT.1)GCTD 996
IF(TUP.GE.1000)GUTO 996
IF(TOP.LE.BCT)GCTC 996
C(TD 997
C(NTINUE
FFINT 998,TDF,BOT
FCRVAT(* TOF=*,18,* HDT=*,IB)
STOP ”PLAGOCLASE ERROR T..BAD TOP/EDT"
CONTINUE
NEUT=FOI
LIFIT = O
CFSSTD(1,1) = CPSSTD(Z,2) 8 CPSSTDC353) 8 0.0
PRINT *,” HILL BACKGROUND CORRECTIONS PE REQUIRED? Y CR N"
READ 57,19
FCRMAT(A1)
ITEST 3 18 .FG. ICHECK
DD 50 181,3,1
FRINT 2,ISTD(I)
FORMAT(I10X,” ENTER CFS FOR ",AZ," STANDARD.")
00 60 K81,3,1
IF(I .NE. K) 6010 ‘0
IF( ITEST) GOTO 901
FFINT 3,EL(K)
FORMPT(1X,A2,“=")
READ *,CPSSTD(I,K)

GCTC 902
CONTINUE
PRINT 30,EL(K)
EORMAT(TX,A2,“ BACKGROUND VALUES ARE ")
EKG 8 RVG("BACKGRD"9
FRINT 31,EL(K)
FUENAT(1¥,AZ," PEAK VALUES APE")
PEAK = AVG("PEPVVAL")
CFSSTDCI,K) 3 PEAK - EKG
PRINT 32,CP$STD(I,K)
FURMATCTOX,"CDRRECTED INTENSITY = ",FTO.Z/)
CONTINUE
(UNTINUE
CLNTINUE

DHINZ : DMIN? = DKINA = 1.05300

DHAxZ = b‘AxE = DNAxL a -1.CEZCC

URITO(10,111) ID,0AIE(I),TIME(1),chSTD(1,1),CPS$TC(Z,2)

FCRVAT(" SAfiFLE :”,2A10,10x,:1o,1ox,a10/”0"," :N STANDARD = ",F

+8
1.2,“ AB STANDARD 3 ",Fé.2/I)

xscousrxttoVN1LTSP))

1000

70

80

85

£231

197

198

59

CONTINUE
XFCI=DYHTCNDUT)
CFS(1)thYN3(NOUT)
CfS(2)=DYflé(NDUT)
CFS(3)=0.0
RQUT‘NDUTRT
SUV = 0.0
CO 70 1=1,3,1
If(CFSSTD(I,I).NE.0.0)C(I) 8 (95(1) I( CPSSTDCI,I> * EETA(1))
15(CPSSTCCI,I) .EQ. 0.0) C(I)=O
SUM = SUP + C(I)
CONTINUE
SU? = SUV * 0.0“
DL 80 1:1,3,1
8(1) = CC!) I SUM
CCNTINUE
LIMIT 8 LIMIT * 1
IF(LI%IT .GT. 1000) CALL SYSTEH(52,” -T00 HUCP DATA")>
DYNTCLI‘IT) = )KCI
DYLZCLIMIT) U(1)
DYNBCLIFIT) HCZ)
DYNLCLINIT) = UC3)
UFITE(10,85)(ISTD(K),V(K),K81,3),XMCI
DHINZ 8 AHIH1(DYN2(LIHIT),DMINZ) S DVINS = AKIN1(DYN3(LINIT),CMIN3
+)
CHINA 2 AMIk1(PYkL(LIHIT),DFIN4) 3 DMAXZ 8 AFAX1(DYN2(LIKIT),D%AX2
+)
DHAX3 = AHAX1(DYN3(LIMIT),DMA¥3) S DKAXG 8 AKAX1(DYN4(LIMIT),DMAXG
0)
FDRKITCTX,“NCRPALIZED FLAGIOCLASE IS ”,3(AZ,1X,F6.2,1X),” MICPDNS
+APE ",F7.2)
IF(LOUT.LE.TCF)CDTG 1000
Y=12.0
N=LIFIT
PRINT 4231
TCRNFT(* DO YOU WANT TO SEE THE PLAGTHCLASE PLOTt)
CtLL VETGCTOKEh,TCKEN2,XNUM)
IFCTCrEU.EO.1HN)PETUFh -
PRINT *,"ENTER PLOT HANTED..1=AN,2=AB,3=OR”
READ *,ITYP£
CALL PLOTSCPLOTBUF,1027,D)
CALL TACTOP(.S)
CALL PLIHITC45.0)
DYN1CLINIT91) 8 0.0
DYL1CLTKIT02) 8 DYLTCLIMIT) I X
CALL PLCTC1.0,1.0,-3)
CALL AllS(0.0,0.0,”h1CRCLS",-7,X,0.0,DYN1(LIDIT+1),DYNTCLIMIT+2))
GCTC (197,198,199), ITYFE
DYNZCLIMIT+1) = DMIN?
DYN2(LIMIT‘2) = (DHAXZ - DFINZ) I Y
CALL AXISCO.D,0.0,"FERCENT AN",10,Y,90.,DYNZ(LIKIT+1),DYN2(LI"1T*2
+))
CALL LINECDYN1,DYN2,LIMIT,1,0,0)
CALL NEVFEKCI)
GOTO 33333
DYN3CLIVIT+1) = DETNB
DTNSCLI!1T*2) = (DHAXB --DHIN3) I V
CALL AXIS(0.0,0.0,"PEPCEIT AE",1O,Y,9O.,DYN3(LIFIT01),DYN3CLIMIT*Z
0))
CALL LINE(DYN1,DYN3,LIRIT,1,0,0)
CALL NEUPERCB)

60

GCTU 33333
199 07N4(LIXIT11) =0PINL
DYNCCLIMIT+Z) 8 (DKAIL - DMINC) I 7
CALL AXIS(0.0,0.C,”FERCELT OR”,10,Y,90.,DYNA(LIPIT*1),DYNL(LIKIT§Z
+))
CALL LINFCDYNT,DYN4,LIMIT,1,0,0)
CALL NERFEN(3)
33333 CALL FLCT(0.,0.0,999)
CALL RENARKC" leISHED-1”)
END
FUNCTION AVG(nOTE)
LOGICAL FINISH,IOCA,IODYN2
SUN = 0.0
KfUNT = -1
‘thSH - .TALSE.
904 (UNTINUE
IF(FINISH)COT0 935
PRINT 10,NOTE
1O FORMATC" hEED VALUE FOR ",A7l)
READ *,TEFF
TINISh = TEFF .50. 0.0
SUV 8 SUE 1 TEN?
KCU”T 8 KDUAT 1 1
60T0 904
905 C(KTINUE
AVG=SUNIFLCAT(¥CUHT)
IACKCUNT .E9. 0) AVG=0.0
RETURN
EKD
SUFROUTINE STOFE —
COMRON/IA,CYN1(1002),rYN2(1002),0YN3(1002),0Y~4(1002)
CCVHLN/TCEIFCFC30),FFLCC)
CALL RETUPNFCSLTAEE7)
N=(N/2)*2
Ika.GE.S)GGTD 30
PAINT 2000,R
2000 TCRhAT(* N=*,16)
STCF "STCRE ERROR 3..T£D FER POINTS”
FFINT 1000
READ 1100,rr~(1)
1000 EORHAT(* FILTER 0R FOURIER?*)
1100 FDR?ATCA10)
'FFIkT'900
900 FORMATC8 ENTER SAMPLE-8)
READ 890,rrn(2)
899 FFR'ATCATO)
PAINT 901
901 EGRAAT(* FNTEc (CAIN-*)
PCAD 899,rru(3)
FRINT.9OZ
902 FCRVATCO ENTER TRACE°*)
qun 999,9rutc)
CALL FFFDE(SLTAFE7,FFN,F06,20)
CLLL PFPAACFDR,“RP“,300)
PRINT 3000
3000 FDRRAT(* STnRING THE DATA.....8)
PUFFER 001(7,1)(~,LYN4(1001))
X=PFCATCTDE)
16(r.£&.0.0)6510 b0
CALL CKPFERR(X,0)
STOP ”STGRE ERROR b..8AD CATALOG”
£0 CONTINUE
-¢RINT-£000,N~
£000 FCRMAT(* *,IS,* PCINTS STOREDA)
RETURN
rfflo

BIBLIOGRAPHY

6T

BIBLIOGRAPHY

Bence, A. E., and Albee, A. L., 1968, Empirical correction factors
for the electron microanalysis of silicates and oxides:
Jour. Geol., v. 76, p. 382-403.

Bottinga, Y., Kudo, A., and Neill, D., 1966, Some observations on
oscillatory zoning in plagioclase: Am. Mineralogist, v. 51,
p. 792-806.

Bowen, N. L., 1913, The melting phenomena of plagioclase feldspars:
Am. Jour. Sci., 4th ser., v. 35, p. 577-599.

, 1928, The Evolution of Igneous Rocks: Princeton, N.J.,
Princeton Univ. Press, 332p.

Cahn, J. N., 1960, Theory of crystal growth and interface motion in
crystalline material: Acta Metallurgica, v. 8, p. 554-562.

Carmichael, I. S., Turner, F. J., Verhoogen, J., 1974, Igneous
Petrology, New York, McGraw-Hill Book Company, 739p.

Chalmers, Bruce, 1964, Principles of Solidification: New York,
John Wiley & Sons, 319p.

Harloff, C., 1927, Zonal structure in plagioclase: Leidsche Geol.
Mededee1., v. 2, p. 99-114.

Hydmann, D. R., 1972, Petrology of Igneous and Metamorphic Rocks:
New York, McGraw-Hill Book Company, 533p.

Klein, L., and Uhlmann, D. R., 1974, Crystallization behavior of
anorthite: Jour. Geophys. Research, v. 79, p. 4869-4874.

Lofgren, 6., 1974a, Temperature induced zoning in synthetic plagio-
clase feldspar, in Mackenzie, u. S., and Zussman, J., eds.,
The Feldspars: Manchester, England, Manchester Univ. Press,
p. 362-376.

Rutter, J. R., and Chalmers, B., 1953, A prismatic substructure

formed during the solidification of metal: Canadian Jour.,
V. 3], p. 15-39.

62

63

- Sibley, D., Vogel, T., Walker, 8., Byerly, 6., 1976, The origin of
oscillatory zoning in plagioclase: A diffusion and growth
controlled model: Am. Jour. Sci., v. 276, p. 275-284.

Smith, J. V., 1974, Feldspar Minerals Vol. 2, "Chemical and Textural
Properties," New York, Springer-Verlag, 690p.

Vance, J. A., 1962, Zoning in igneous plagioclase: Normal and
oscillatory zoning: Am. Jour. Sci., v. 260, p. 746-760.

Neibe, R. A., 1968, A record of magmatic conditions and events in a
granite stock: Am. Jour. Sci., v. 266, p. 690-703.

PLATES

64

65

.QQIN a

.AE: ompv

.Ae= opmv .vulm omgm>mgh mew; "N qumxgu muwumA
.AE: ommv .mm-~ mmgm>mgp mew; "N Pmumxgo muwamU
.uoum ..mm-m mmmgm>mgh mew; "N pmumxgu mpvqu
.As: ompv .ouup mmgm>mgh mcw4 ”_ Fmpmxgu mu_UMA
.nn1F ..mm-_ mmmgm>mgp mew; up Fmpmxgo mappmA
.nnup ..mm-— mmmgm>mgh mew; up ngmxgu muwumJ

o .UUIF .

.umpo: mmwzgmcuo
ogmsz pamuxm measure ash Fosao mcucw: ouoga _F<

_ mumpa

Nugmzo
Nugmao
Npgmao
~uLm30
Nagmzo

~pgm=o

mgsmwm

66

 

.o o.u

 

o a:

—

    

Wb<4m

67

..nnum mew>mgp mew; Pmpngu mumumJ

..x¥-m ..Pw-m mmtmsaeh mew; Fmpmxgu ou_um3

.Aeo_mv sp-m ..xx-m ..nw-m ..wp-m amem>~Lp mew; _mpm»gu upwumd

.AE: omeV noun mmgm>mgp mew; qumxgu mawumJ

cad-«31.01010

..;;-¢ ..mm-¢ mmmgm>mgp mew; Fmpmxgo muwumA

..Lw-m mmgm>mLp mcw4 Pmumxgu maven;

.umuo: mmwzgmcuo
ogmzz ugmuxm mcoguPs cum szcm mguUwz oposa PF<

N muupm

Nugmao
Nugmao
~pgmao
upsoao
Nugmzo

Npgmso

LO

NMQ‘LD

mgzmwm

68

 

c .2“.

N uh<4m

 

 

=10

_ 6:.

69

.AE: ompv .wmuw mmgm>ogp mew; m
.As: ommv 0:16 .suu¢ .Lmue ..mm-e mmmgm>mgh mew; "e qumxgu mpwunvozsm e
.ao1m ._mm-m mmgm>mgh map; "m Fmpmxgu mumumuoxgm m
.cnum ..uu-m ..uu-~ mmgm>mgh mew; "N Pmumxgu mpwomuozzm N
.Ae: ommv .anup ..mm-P mmgm>mgk ocw4 "_ qumxgu mpwumcoxnm F
mgamvu

.umuoc mmwzgmsuo
mgmsz pamuxm mcoguws cum Fmaam mcpuwz opoca —F<

m mumpm

70

 

n mh<ga

0016

a¢-n

n .o.u

 

71

.A53 ommv mu-m ..Fp-m ..xx-m mmmcm>mgk mam; "w Penmagu mpwumuo»;m v
.Asz ommv .nn-n mmgm>mgh mew; an Fmpmxgu muwomvoazm m
..Fw-o ..;;-m mmtm>mLh m=_4 no FmUJALu muvuauoagm N
.mmum mmgm>mgh mew; "m pmumagu mpwumuo>5m _
mgzmwm

.umuo: mmvzgmguo
mgmgz pamuxm measure on“ Peace mspuwz ouogn FF<

e «papa

72

 

v.9u

.5
\

ma;

:5

 

v

 

To:

wh<qu

 

73

.nx-~ mmgm>mgp «cw; ”N Pmpmxgo ppmmmm
.nmnm mmgm>ogh mew; "m qumxgu upmmmm

.AE: ommv Naup .oxup mmmgm>mL» mew; up Pmumago upmmmm

.umpo: mmwzgmsuo
mcmsz pamoxm mcogu_s cum szcm mguuPz ouoga ~F<

m vamp;

mgzmwm

74

 

n uh<4u

“'11111111[11111111111111‘3