Download ARCHEAN METAVOLCANIC ROCKS FROM THACKENAY

Canadian Mineraloglst
Vol. 12, pp. 509-519 (1974)
ARCHEAN METAVOLCANIC ROCKS FROM
THACKENAY TOWNSHIP. ONTARIO
T. H. PEARCE eNp T. C. BIRKETT
Departmentol GeologicalSciences,
Queen'sUniversity,Kingston, Ontario, Canada.K7L 3N6
township is Lat., 48o27lN, Long., 8o"01{il}The rocks in this area form part of the KenojeThe iron-rich Archean submarine metabasaltsof vis Group @. Pyke, pers. commun. 1974). This
Thackeray Township, Ontario, are thought to repre- area was selected for detailed study after discussent a distinctiv: magma type: high-iron basalt.
High-iron basalt is defined as having )t2Vo by sion with L. Jensen, who has mapped several
\peight FeO (total iron as FeO in the anhydrous townships, -including Thackeray, at a scale 1:
tA mile) for the Ontario Division
analysis). Two well-exposedflows, one of which is I584O (I'
a typical high-iron basalt, have been studied in of Mines, (Jensen1969, l97la, L97Ib, 1973).
detail. The metamorphism of the region is preh- The area has: good outcrop exposure, a subnite-pumpellyite erade yet both flows have original greenschistgrade of regional metamorphism, pdtexturesand mineralspreserved.The chemicalcom- mary textures and minerals, no structural composition of the massiveparts of the flows appears plications such as penetrative
deformations or
to be unchangedby the metamorphism,exceptfor
folding
on
the
of
scale
an
outcrop,
and good acH2O and COz. Plagioclasewas the inferred liquidus
phase in bctth lavas. Pyroxene, hornblende,and cess by logging roads. Accordingly, this was
Fe/Ti oxide grew during post-eruptive crystalliza- judged an excellent area for detailed study of
tion. The presenceof skeletaland frond-like horn- one of the important rock types comprising Arblende in the upper part of the iron-rich flow sug- chean greenstonebelts.
geststhat hornblendemay have grown as a quench
The basaltsof Thackeray Township lie above,
phase,and was probably metastable.The differentiation of the flows car be modelled by crystal and are probably part of, a 20,000' (6100 m)
fractionationin which amphibolefractionationplays series of basalts in Garrison Township to the
a key role.
north. Goodwin (1,972)rsports an average analysis of seven samplesfrom unspecified locations
in Garrison Township which were reported to
cover the entire section of 6100 m. and this
INrnoouctroN
is similar to our results.
The metavolcanic and metasedimentary
greenstonebelts of the Archean shields are of
Frst-p DescnlprloN
great theoretical and economic importance. Detailed chemical and petrographic descriptions of
A general description of the rocks of Thackthe basaltic rocks which are common building
eray Township is provided by Jensen(1973). The
blocls of the greenstone belts are necessary for
are
the interpretation of the Earth's early history. In rocks in the northern part of the township
-@ig.
this work, we describe a rock type from the a series of pillowed and massive basalts
Abitibi greenstone belt, 'high-iron' basalt. We 1a), intercalated with thin chert layers and some
define high-iron basalts as rocks of generally thin volcaniclastic units (Fig. 1c). The basalts
basaltic composition (45-52% SiOa) containing are typically dark green, fine- to medium-grained,
)I2Vo by weight total iron expressedas FeO hard, and have a conchoidal fracture. Outcrops
calculated on an anhyrous basis. This magma are generally rounded and of low relief. The
type appears to have occrured as subaqueous, massive, coarser-grained, centers of the flows
hydrous, lava flows which were tlick enough to tgn{ to weattrer high and commonly display a
have differentiated during solidification. We distinctive texture whicho for want of another
have since recognized this rock type in other term, and to draw attention to its significance,
greenstone belts of the Superior Province of the we call "reverse-diabasid's.In this texture. latls
of euhedral hornblende are set in a matrix of
Canadian Shield.
plagioclase. Close scrutiny of the weathered
The rocks of this study are located in northern surface is necessary to distinguish this texsrre
Thackeray Township, in the central part of the
Abitibi Greenstone Belt (the NE corner of the tsee footnote on p. 510
ABsrRAcr
509
510
THE CANADIAN MINERALOGIST
from normal ciiabasic texture (euhedral plagioclase set in a matrix of pyroxene-see flg. it).
The fine-grained flow margins are commonly
exposedat the edgesof outcrops and .may easily
be missed in routine field work. The hieh-iron
basalts locally exhibit skeletal quench hornblende in the upper parts of the flows, and in
pillows.
Minor veins of two ages cut the basalt: an
older set containing quariz and epidote, and
having diffuse boundaries, and a younger set
containing quartz, calcite and minor epidote
with sharp boundaries. Meta-domain. ui d"sqibed by Smith (1968)and Jolly & Smith (1972)
are not developed except as thin selvagesalong
joints and near quartz-bearing faults. D,eforma-
tion on the outcrop scale (-3Om) has apparenfly
been accomplished by movement on widely
spaced, discontinuous fractures now lined with
chlorite. Veins of albite are present locally, especially in breccia units (see Fig. 1c).
'rWe call this textur: "reverse
diabasic" because
the role of the dark and light minerals are reversed
(compared with a diabasic texture). The dark
mineral forns euhedral prisms surrounded by the
light feldspar. lt is important to distinguish these
two textures because the one is usually present in
intrusions such as dykes, while the other is crmmonly exhibited by Archean submarine flows. We
do not yet know how common this ..rev:rse diabasic" texture is in either Precambrian or phanerozoic rocks.
*:,;..,'.'i:;
FIc. la: Pillow basalt 30 m below Flow 2,
Ftc. 1b: Reverse-diabasictextures on an outcrop 10 m below Flow l.
Frc. 1c: Albite vein cutting volcaniclasticunit immediatelv below Flow 2.
Flc. -1d: Altered plagioclase phenocrysts interpreted to have floated up in Flow 2. White veins are
considered .to be the last liquid to crystallize in the flow. The hammer handle points towarJs the top
of the flow.
METAVOLCANIC
Fre.
Frc.
Frc.
Ftc.
2a:
2b:
2c:
2d:
ROCKS
FROM
THACKERAY
TOWNS,HIP.
ONTARIO
Radiating amphibole in a matrix of amphibole and plagioclase, Flow 1.
Contact between shards and spherulites at base of Flow 1.
Quartz-prehnite-actinolite-epidote-chlorite-magnetite-pumpellyitevein in chert below Flow 2.
Swallowtail quench plagioclase in the centre of a spherulite, Flow 1.
511
SKIN
,,(
,|(
+t-
VESICLES
c>*
RADIATING
AMPHIBOLE
GRANULAR
BASALT
a
a"
at
o
a
aa'a
o.'
ro!l,o'r
li:.i:.':i CUMULATE
a
taa
a
a
aa
ao'
SPHERULITES
S H A RD S
Columnar Section of Flow 1.
Frc. 3a: Spherulites with radially grown chlorite forming an interference mosaic matrix. Flow l.
FrG. 3b:_ Microporphyritic pyroxene and amphibole with a matrix of hornblende,plagioclaseand.quanz,
Flow 1 cumulate zone.
Frc. 3c: Skeletal radiating amphibole, Flow 1.
Frc. 3d: Amygdule, formerly a segregationvesicle, Flow 1.
Frc. 3e: Concentrically zoned shards in flow-top skin, Flow 1.
Note: that the thicknessof the zones in tle column have been exaggerated.See text for the true thicknesses.
METAVOLCANIC
ROCKS
FROIVT TIIACKERAY
MlcRoscoPrc DrscrupnoN
The first of two flows selected for study is
a 30- to 35-meter thick flow of high-iron basalt
exposed near the main logging road. The base
of the flow is a 1- to 2-cm thick layer of flattened structures which we interpret as former
glass shards (Fig. 2.b). They are now chloritized.
Above this layer, there is a 10-cm zone of
spherulites (Fig. 3a). These spherulitesare radial
intergrowths of plagioclaseand a mafic mineral
which have nucleated on skeletal quench plagioclase microlites (Fig. 2d). The spherulites now
occur in a matrix of chlorite which has grown
radially from the spherulites into a polygonal
mosaic of mutually interfering domains, (Fig.
26, 3a). Overlying the spherulitic zone is a
10-m microporphyritic layer @ig. 3b) containing
microphenocrystsof hornblende, some of which
rnay be in part pseudomorphousafter pyroxene,
since they have pyroxene cores. Skeletal magnetite occurs locally.
The microporphyritic zone grades upward into a zone of granular hornblende basalt approximately 10 meters thick which, in turn, grades
upward into a zone containing radiating hornblende (Fig. 2a), locally skeletal (Fig. 3c). We
interpret this as a quench product'F.The quench
texture appears at its lowest point in the flow
as clusters of stubby, radiating hornblende
crystals in an otherwiseequigranular hornblendeplagioclasebasalt, and reaches full development
near the top of the flow as sets of frond-like
growths of crystals several millimeters long,
intergrown with plagioclase,and including some
skeletal plagioclase microlites. No preferred
orientation of the hornblende or plagioclasewas
observed.
Near the flow-top, composite amygdules similar to th6se described by Smith (1967) as segregation vesicles make up a few per cent of the
rock (Fig. 3d). The segregation vesicles have
sharply defined inner surfaces resulting from
the re-entry of lava into the vesicles. Surrounding
each vesicle is a darker zone with gradational
outer boundaries.The segregationvesiclesoccut
in a fine-grained radiating amphibole rock, and
typically contain late fillings of quartz, chlorite,
epidote, and sulphides, with occasional actinolite.
The flow is capped by a thin skin of ropy
lava containing former glass fragments (Fig.
3e) which shows the effects of repeated fracturing as the flow moy€d. The shards of broken
*By quench product we mean crystals gtown from
t1le liquid during relatively rapid cooling of the
flow. The grorrth rate is relatively rapid for the
mineral species involved so that the typical form is
skeletal or feathery.
TOWNSHIP.
ONTARIO
513
glass contain skeletal plagioclase microphenocrysts. Like all of the finest-grainedrocks in the
study area, the flow top is veined with quartz
on a fine scale, and is thoroughly chloritized.
The second flow selected for detailed study
lies in the noftheasternpart of Thackeray Township. The lower 65 m of the flow are exposed
and it is not high-iron basalt by our definition.
The base of the flow is a fine-grained, chloritized, basalt.Above the baseis a microporphyritic
layer, containing mostly pyroxene, with some
horn.blende,and amphibole of deuteric origin.
Plagioclase phenocrysts up to 3 cm in size iucrease in amount upward in the flow, then end
abruptly in a thick layer (Fig. 1d). Flow 2 contains no radiating amphibole in the exposed
part.
INrunpnererroN
oF Trxruner
VeruerroNs
We interpret the textural features of Flow I
as follows. The shards at the base are the chloritized skin or rubble formed as the flow was
moving, and were flattened by the weight of the
overlying flow. The spherulitic layer immediately above the base is due to devitrification of
the glassybottom of the flow after the flow had
stopped moving (since the sph,erulitesare not
deformed). The skeletal microlites of plagioclase
are quench crystals formed during post-eruptive
crystallization of the flow. The amphibole and
pyroxene phenocrysts of the microporphyritic
zone are cumulate minerals formed while the
flow slowly cooled, and which have settled
towards the base. There is some evidence of
magmatic reaction between the pyroxene and
amphibole, but we ile not sure of its extent.
The radiating amphibole fronds associatedwith
quench plagioclaseare quench crystals. We feel
they cannot be metamorphic minerals because
they show definite stratigraphic variations (in
size and amount) within the flow, and they are
not present in some of the nearby flows which
have the same grain size and mineralogy, and
have presumably undergone the same metamorphism. The alteration of the glassy parts
of the flow may have begun during the eruptive
phase by reaction with sea water, but this is
difficult to prove.
Flow 2 is similar to Flow I except that it
lacks quench amphibole, implying, we feel, a
lower water content at eruption. It atso has a
layer of large plagioclase phenocrysts, which
we interpret as due to crystal fractionation by
flotation. The upper contact of the flow is not
exposed, and the features at the base are Dot
clear because tle base is intiniately assoaiated
with a 15-m thick underlying fragmental unit.
51.4
THE
CANADIAN
MINERALOGIST
TABLE1' WHoLE
RocKANALYSES
FRoMTHACKEMY
Tol'lNsHIP,s0MEPUBLISHED
ANALYSEs
F0RC0MPARIsoN,
pRSBE
ANDELE6IRS1{
ANALysEs
0F
123456789]01.1
si02
4 7 . 1 0 49.9q
' 'v2
u.vJ
-g . g i
Al z0r
14.45
1!.70
Fe;oi 2.23
-\ .77
leu
12.66 1 0 . 0 6
Mno; 0.22 g.?0
Mgo
B .1 5 9 . 2 8
Cao
1 3 . 4 3 1 1. 7 7
Nad
?.00 9 . 9 8
KZo
0.23 0 . 4 7
PZlS
0 . 1 9 0 .1 3
suM 101.59 98.63
ilelght*
O
3,4
47.00
0.83
14.38
1. 6 6
9.39
o.le
9. 0 2
l l .90
1.63
0.32
0 .1 4
96.46
7.A
47.10
0.78
14.00
1. 7 1
9.69
0.18
9.85
l '' l,.t0. 5
30
47.70
0.75
14.43
I .58
8.95
0.18
9.86
11. 8 0
1.37
0.34 0.28
0.12 0..13
96.12 97.03
7.4
LL.1
rerzrsr9z1t
Si0r
I r u 2'
At 203
5l.t0
1.42
13 . 6 0
2
.04
3
Fe0
II.56
Mn0
0.19
r'190
7.00
Ca0
9.34
Na20
2.02
K2d
0. 6 8
Pzos 0 . 2 2
sUM
99.17
lleighlt*
J.J
49.80
1.72
12.90
2.14
12.t5
0.22
6.B0
9.40
I.80
0.61
0.25
97.79
1.5
48.30
0.73
1. 4t . 7 5
.40
7.95
0.07
8.90
12.04
1.20
0.32
o . ,tl
95.77
17.8
48.20
2.65
10.50
2..52
I5.98
0.25
6 .I 5
6.32
I.90
0 . . I1
0.3B
95.26
50.50 51.20
1.75 I .Bl
I 2.50 12.95
' 122. .1023 2 . 3 1
13.09
0.19 0.20
5.95 6.70
8.40 8.82
2.10 2.10
0.65 0.55
0.28 0.28
96.47 t0o.o.t
50.30
1.82
12 . 3 0
2.15
12.20
0.20
6.60
8.75
1.45
0.55
0.27
96.59
14.8
20.4
29.3
25.2
46.50 47.@
0.73 0.69
14.03 15.30
1 . 6 6 1. 4 4
9. 4 4 8 . 1 3
0.t9 0.16
5 9.80
' l9l ., 32 5
II.96
o.75 1.?3
0.21 0.29
0.15 0.13
94.26 96.V3
2 1 .L
22
5' 0I . 0 0
.70
12.60
2.23
12,66
0.20
6.95
9
' 1. 5 5
,68
0.36
0.20
98..13
1.1
23.3
12
46.90 49.30 52.00 4 7 . 4 0
0 . 7 3 0 . 7 4 0. 7 4 0 . 8 0
1' l5 , 4 0 1 6 . 5 0 . 1 5 . . 1 01 4 . 3 0
'1.73
.54
1. 4 2
I .63
8.76 8.03 9.24 9 . 8 2
0.17 0.14 o.l7 0.19
8 . 2 5 7 , 4 5 8 . 4 2 8.80
1 2 . 3 8 9 . 9 0 1 0 . 8 8 10.77
l.40 .1.70 1.82 0.09
0 . 3 7 0 . 3 9 1. 0 2 0 . 9 3
0.17 0.19 0.15 0.13
9 6. 0 7 9 s, 7 6 1 0 1, 1 7 9 4. 9 6
28.b 32.2 34.4 40.?
24
23
50.05
I.58
I 3.20
2.05
II.64
0 .1 9
6.72
8.75
2.48
0.57
0.25
97.qB
4 9 . 0 0 5 t . 9 0 49.34
' 112. .9075 1 , 6 7 ' 119. .4094
1 2 . 7 0 'l
.99
' I22. 1
. 14I 2 , 0 4
1 1 . 5 6 6.82
0.25 0.22 0.17
5.97 5.87 7.19
9 . 9 0 8 . 9 8 t 1. 7 2
2.W
2.95 2.73
0 .. 4
'0
3 21 0 . 3 7 0 . t 6
0 . 2 4 0 .1 6
96,sz 98.50 9B.81
1L.9
34.8
zs
26
nn.22
2.03
13.62
?.21
12.36
0.26
4.36
8.47
3.26
0.36
0.35
9 7. 5 1
't4
13
15
47.10 47.m 49.50
0 . 6 6 0 . 7 0 0. 7 1
1 4 . 6 0 ' 1 3 . 9 0 7. 4 0
.1..43 I .59 1.26
8 . 12
9 . 0 4 7. 1 4
0 . 16
0.18 0.22
7 . 5 8 ' 182. 0. 1
02 I3.70
12.77
12.65
2.10 r.00 1.72
0.59 0.45 1.70
0.t4 0.14 0.20
9 5. 2 5 9 4. 3 2 9 6 . 2 0
41,b 48.5 49,6
ZB-
-
d---
48.50 51.16
0.48 0.36
9.60 3.81
' 127. 2. 158
22.51
0.57 0.61
8.63 8.64
r0.20 n,52
0.66 0.61
0 . 0 8 0 . 1I
0.00
9 8 . 15 9 9. 3 3
30
49.91
0. 4 3
5.30
18.47_
0 . 4 1t
i1.09
11.68
0.71
0 . 11
9 9. 0 1
2 7. 8
FL0l,l2-rocksfrcmtieT.73series:ana1ysls(l)T-73-93,(2)?9.:(3]-97,-(4)9s,(5)99m
( e ) r 0 4 ,( l o ) l o 5 ,( n ) r o z ,t i 2 j - i o g (; r c 1 ' i r o l ' t r cl lt t ,
.
t h eT - 7 3 ' s e r l e(si o: 1 ' r - z r l t ) '( r i j - i a ) t i a j ' i i , 'trli56)l ni 2
i , ( 2 0 )3 3 , ( 2 r ) 3 4 ,( z z ) 3 5 ,( 2 3 \ 3 8 ,( 2 4 ) 4 2 ,
"ott
filf'ufrom
az're65)' !:i?ffii,i:'llT.fi,;iH.offl,ll:.'*
&shaw
,e74).
i?ii
IiiJ
( 3 0 ) fiiili;r:",:3:l',fl;:::iil;n'?;l:1,1fi?T.l'3erFe'
Q u e n c ha m p h i b o l ef r c m a h . i g h - i r c n b a s a l t .
*Height in meters
a D o v et h e f l o h b a s e .
Cnetvrrsrny
Analyses of the two Thackeray basalt flows
are presented in Table 1. All analyses were
produced by XRF techniquesexcept MnO which
was done by atomic absorption. Because of
FLOW I
o
FLOW 2
'
O T H E R T H A C K E R A YB A S A L T S O
Frc. 4; AFM -diagram for
Thackeray Township.
meta-basalts from
variable oxidation in the coarser parts of flows,
all analyseshave had their iron content adjusted
to make FerOs : 1,5Voof total iron as FeO in
weight units. This is the Fe2+/Fe3+ ratio for
the quench amphibole rocks which are thought
to be the least altered.
Flow 1 is a high-iron basalt similar to other
basalts in Thackeray Township (other analyses
will be presentedat a later date), and similar to
the high-iron basalts of Garrison Township to
the north (Goodwin L972). These rocks probably
represent a single high-iron basalt series(trig. a)These basalts are broadly similar to the osean
tholeiite of Engel et al. (1965),seeTable 1 #26,
but they have substantially more iron and less
AlrOs and CaO. This rock type closely resembles
the diabasedykes which form major swarms in
the Canadian Shield (Fratta & Shaw 1.974),see
Table 1 ft27. We have .found similar high-iron
basaltsin other greenstonebelts in the Su.perior
Province, for example,Pickle Crow. It may be,
therefore, that this magma type was a fairly
common product of Archean volcanism and
possibly characteristic of certain stages in the
growth of the volcanic edifices. More work remains to be done on this subject. Flow 2 is
not a high-iron basalt by our definition; it more
closely resemblesa typical tholeiite.
At the present time we do not have a good
explanation for the high iron content of these
METAVOLCANIC
ROCKS
FROM
THACKERAY
TOWNSI{IP,
ONTARIO
515
tion, is shown in Figure 5. The amphibole which
we have used is a cumulate amphibole analysed
by electron microprobe (Table 1, no. 28). Point
P in Figure 5 is the assumedparent composition
of the flow - a well-preserved quench amphibole rock (#2O in Table 1). Point D is the
model differentiateo the parent wth 3O% by
Dtrpen'eNtlAttoN
weight of hornblende removed. Point C is the
A fractionation model has been developed model cumulate rock, the parent wilh 3O% of
based on our interpretation of the textures and hornblende added. Addition of about 40Vo
tested using the chemistry of the flows. For hornblende to the parent adequately accomFlow 1 we postulate a homogeneous magma plishes the most extreme chemical differentiaerupted as a thick flow on the ocean floor. The tion observed(#22 in Table 1). Deviations from
'on the molecular
liquidus phase under the conditions of eruption simple linear relationships
was plagioclase which grew as microlites on ratio diagrams are probably the result of rninor
quenching of the margins of the flow. As the pyroxene fractionation, and minor movement
flow cooled further, conditions of temperature, of plagioclase crystals enclosed in cumulate
pressure, and composition (especially water amphibole or pyroxene. Becauseof the generalcontent) were such that amphibole formed as ly good fit of the model to the data we feel that
variations and errors in PgOshave not signifiquench crystals; these are fine near the flow
top, but become coarser due to slower growth cantly affected the results.
lower in the flow. Within the still liquid part
MOLECULAR RATIO OIAGRAMSFLOW t
of the flow, more equant amphibole, and some
pyroxene, formed and settled to the bottom, giving a depleted zone near the centre of the flow.
Tt o2
and a cumulate layer near the base (the micro"2o5
porphyritic zone). We are not cetrain at this
point of our modelling what proportion of pyN o zO
roxene was originally present in the cumulate, .6-.-_
'2 'S
but it may have been relatively srnall. At some
stage Fe/Ti oxide formed, and may have settled
at2 o3
-' 2 e 6
to some extent.
We have,testedthis model using the available
chemical data, and the method of graphical
analysis using molecular ratio diagrams developed by Pearce (1968). Interested readers should
Mgo
review this work in order to understand our
Frosreasoningin detail.
Briefly, in a series of rocks related by crystal
fractionation. if the molar amounts of two
chemical speciesin a seriesof chemical analyses
are divided by the molar amount of a species ._
co O
'z
not involved in the fraotionation (a constant
"a
species),and the resulting points are plotted on
a diagram, the diagram is a molecular ratio
o
diagram. The slope of the trend for any pair
too
of speciesis the molar ratio of those speciesin
F 6 O+ F e Z0 3
----;-=the fractionated mineral or mineral mixture.
'2-6
Becausethere is no evidence for the crystallization of any phosphorus-bearingmineral before the crystallization of the Thackeray flows
aio2 / P2o6
was well-advanced, we posfulate that phosphorus
was a constant element during the fractionation
diagrams for Flow 1. Point
of the Thackeray flows, and can be used to Frc. 5: Molecular ratio
"P" is the model parent composition, 'oC" is the
normalize the molecular ratio diagrams for the
model cumulate composition, and "D' is the
flows.
model ditferentiate composition. Points lying to
Our analysisof the data and our fractionation
the right of "P" are cumulate rocks, those to tbe
model, assumingdominant amphibole fractionaleft are depleted rocks in the flows.
Archean magmas. Modelling work which will
be presented at a later date appears to indicate
that this is probably a feature of the original
primitive magma rather than a result of differentiation.
5r6
THE
XOL€CULARNAlIO
26l
1to2
-F 2 v 6
^l
P2o 5
O I A O P A T SF L O W 2
.
5oT
Nq2o
:---
CANADIAN
I
200
At?o3
-'2"
6
MINERALOGIST
are not involved in the fractionation process
(mainly the so-called incompatible elements)are
"increased' or "decieased", while the major
elements involved in the fractionation show
"less" change in simple weight comparisons.
The net result is that weight per ceDt comparisons do not show the fractionation processin an
easily understood mann€r. Molecular ratio
diagrams as used hereo on the other hand, illustrate the process in a more readily comprehensible manner. This flaw in traditional methods of analysis (such as Harker diagrams) has
been documented for the Palisade Sill (Pearce
r97O).
Mntetvtonp.Htstvt
stoz / P2oa
Frc. 6: Molecular ratio diagrams for Thackeray
Flow 2.
The molecular ratio diagrams for Flow 2
(Fig. 6) show more scatter than those for
Flow 1. This difference is believed to be the
result of fractionation. and uneven accumulation.
of pyroxene, amphibole, and plagioclase. Because fractionation of these three phaseswould
have different intersecting trends on several
molecular ratio diagrams, and the relative contribution of each phase has not been evaluated,
no quantitative model has been formulated at
this time for Flow 2.
Cursory examination of the major and minor
element in the analyses of Flow 1 in Table I
shows that, for the most part, tle major constituents such as SiOz, AlgOg, MgO' and CaO
do not appear to vary (in weight 7a) as much
as some of the minor elementssuch as Kso and
PrOs. At first, it might seem paradoxical that
differentiation can apparently affect major elements less than minor or traca elements. This
is a necessaryconsequenceof crystal fractionation involving a phase whose major element
chemistry closely resembles that of the parent
liquid (for example, clinopyroxene or amphibole
fractionating from basalt). The elements which
Generalizing from the study of Flows L and 2,
metamorphism in Thackeray Township can be
described as having taken place in three stages
(excluding late quartz-vein formation). The
earliest $tage was the initial deuteric alteration
of the flows, the second stage was prehnite-pumpellyite facies metamorphism, and the third stage
was later, low-pressuremetamorphism related to
the emplacement of the Garrison Syenite. We
emphasize again that the metamorphism of these
rocks is incomplete and that most of the original
igneous mineral species have been locally preserved.
During the initial deuteric alteration the finergrained parts of the flows were partly hydrated,
and probably reacted with sea water. Features
which we infer to have formed at this stage
and that survived to the present include sening
in originally glassy volcanic fragments, quartz
veining on a fine scale in flow marginso and
albitization along fractures, usually in brecciated
units (Fig. 1c). Some chlorite probably grew
directly from the glass at this stage (Fig. 2b),
The second stage of metamonphism was
marked by continued burial, with P-? conditions
reaching those of the prehnite-pumpellyitequartz facies. In this stage the central parts of
the flows were partly hydrated. Plagioclase phenocrysts were partly replaced by prehnite, clinozoisite, calcite, quartz, and possibly calcium
montmorillonite. Pumpellyite formed in pyroxene
crystals. Vesicles contain assemblagesof prehnite, quartz, chlorite, calcite, and albite with
epidote, or pyrite. All assemblagesexisted in
close contact with magnetite and sphene. Muscovite is present in small amount$ in the rocks,
and may have been a stable mineral at tlis stage.
During burial metamorphism, monomineralic
aggregates of calcite, clinozoisite, or albite developed, generally on a microscopic scale, indicating that sodium and calcium (and po*sibly
METAVOLCANIC
ROCKS FROIVI TTIACKERAY TOWNSHIP,
other elements)were able to migrate in the rock
over a distance greater than several millimetres.
Quartz-calcite-chloriteveins probably also formed during this second stage of metamorphism.
During the third stage of metamorphism,
prehnite was restncted .to calcium-rich environments (former plagioclase phenocrysts), leaving
in its place in the groundmass distinctive aggregates of quartz grains (Jolly, personal communication, L974). The prehnite co-exists with
clinozoiaite, quartz, albite, chlorite, muscovite,
and probably calcium montmorillonite. Pumpellyite remained the stable phase replacing pyroxene phenocrysts,while in the groundmassthe
stable metamorphic assemblage is some or all
of quartz, epidote, actinolite, chlorite, albite,
calcite, muscovite.All phasesexist in close proximity with sphene and magnetite.
Worthy of special consideration are two seven-mineral assemblages.
ONTARIO
st7
\uhere z"ro is the number of moles of water
involved in the reaction, and dlz" is the change
in volume of the solid phasesin the reactant
and product assemblages.
Figure 7a illustrates that increasingthe chemical potential of COzrelative to that of HzOsuppressesthe stability fields of pumpellyite and
epidote-actinoliteassemblagesin favour of calFigure 7b, shows that
cite-chloriteassemblages.
decreaging& (g! constant T, Furo l.rcoo)would
havethe sameeffect.
FlG.7o
quartz-chlorite-epidote-actinolite-pumpellyitemagnetite-calcite inferred from observations
of Flow 2.
2. quartz-prehnite-pumpellyite-opidote-actinoliteo
chlorite-magnetite, observed in a single thin
N
E
section from a vein in chert about 100 m
stratiCraphically below Flow 2 (see Fig. 2c).
l.
If we assume that the compositions of all the
minerals are adequately represented in the
model systern SiOz-ALOa-FezOe - (Fe, Mg)
then both assemblages
O-CaO-COa-HzO,
would be isothermally univariant in pcozpgro-P.oilu" space. On any two-dimensional
isothermal section of this space (such as the
isobaric F*r- Lrrro diagram) all seven minerals
of each assemblage could co-exist stably only
at an invariant point. Using the mineral formulae in Table 2, the "reactions" among these
mineral assemblages have been deduced and
balanced. The equilibrium curves of the 7phase assemblagesare shown in Figures 7 and 8.
The slope of an equilibrium curve on a P"puro dia8ram is given bY
lPs
dp^"o
:
d?,'"
dvt
(41.84)
TABLE2. STRUCTUML
AIIOIOLARVOLU}iIES
FORI'IJLAE
OFI'IINERALS
USED
TO
CALCI,ILATE
THE TNVARUNTPOINTSOF FIGURES7
PTPQCCIE
3.00 6,00 1.00
1
.90 4.00 0.00
?|d:
0.10 1.00 0.00
0.00 1.00 0.00
Cao
2
' 1. .0000 4 . 0 0 0 , 0 0
Hfr
3.50 0.00
w2
0.00 0.00 0.00
vGc/mle) l4t.l0 298.00 22.69
5.l02
tiIilSro
AcM
0.00
5 . 4 6 3 . 0 0 8.00 0.00
0.00
5.04 2.25 0.00 0.00
0.00 0.05 0.75 0.00 1.00
0.00 9.46 0.00 5.00 1.00
1.00 0.00 2.00 2.00 0.00
0.00 8.00 0.50 I .00 0.00
l.Q0 0.00 0.00 0.00 0.00
I
36,90 426.30 139.00 280.00 44.52
Pr prdrnlte P pmpellylte Q quartz C ca'lclte
E epldote Ac actimlite
!l nagnetlte
Cl chlorlte
ALL ASSEMBLAGES
T N c L U DeEu A R T z
A N DM A G N E T I T E
coNsrANT 'T
3
/l co2
(c,cl)
Flc.7b
Ps
(A)
C o N S T A N TT ,
<-il
Co,
l. Hzo
Fig. 7a: The invariant point quartz-magaetiteepidote-actinolite- calcite-chlorite-pumpellyite at
constantRand ?.
Fie. 7b: The invariant point of Fieue 7a at constant
T, p'(;c,2.The two diagams represent orttrogonal
planesin Ps- pa2o-peo2spaceat constant ?.
Solid circlesindicate observedassemblages
of metamorphicminerals.
Q-quartz, M-magnetite Bepidote A-actinolire
Gcalcite, Cl-chloritg P-pumpellyite
Jl6
THE
CANADIAN
Figure 8 shows that the prehnite-pumpellyite stability field is relatively unaffected by
changing P(at constant T, Fszofcor) but that
the assemblage chlorite-prehnite has an upper
limit of P that is independent of p-r.
The assemblagesassociated with the late
thermal event are similar to those reported by
Coombs et al. (1970) .for the Northern Appalachians. Since the stratigraphic interpretation
of the Abitibi GreenstoneBelt implies prehnitepumpellyite facies metamorphism at considerable depth because of the great thickness of
volcanic rocks, we infer that there must have
been a break in time between stages 2 and 3,
although the maximum temperatures during
these stagesmay have been quite similar.
(P^r,Cl)
MINERALOGIST
Since prehnite is still present in the Thackeray
rocks, and assuming equilibrium, conditions of
the last thermal metamorphic event could not
have exceeded400'C, the upper thermal sta,bility limit of prehnite (Liou 1971). We feel that
the pressureon the rocks at the time of the last
thermal event may have been quite low, perhaps as low as 2 kb.
Samples selected for chemical analysis were
generally massive rocks, although samples with
small amounts of veins or amygdules were also
analyzed. Despite wide variations in grain size,
and degree of hydration, the rocks within each
flow show coherent trends on molecular ratio
diagrams (Fig. 5, 6). The position of the rocks
in the molecular ratio diagram is consistentwith
the position inferred from our interpretation
of the rock texture. We believe that this indicates that, except for the addition of water,
and some carbon dioxide, the massive rocks of
the flows have suffered little change in their
bulk composition. We do not extend this interpretation to zoned volcanic fragments or to rocks
with large amounts of quartz or quartz-chlorite
veining.
CoNcLustoNs
The two Archean submarine metabasaltflows
from Thackeray Township which have been reported in this work, have been metamorphosed
to prehnite-pumpellyite quartz facies conditions.
Their well-preseryed textures and primary minPr
erals suggest that, apart from COz and HzO,
the central parts of the flows have retained their
original composition. The basalts have a high(e
iron content and we suggestthat this is a magma type which is probably widespread in the
Archean. The flows are about 35 meters thick,
and have differentiated during post-eruptive
sl
crystallization. The flows crystallized skeletal
plagioclase (the inferred liquidus phase), hornCONSTANT T
blende, and Fe-Ti oxide. The presenceof skeletal
and frond-like hornblende in the upper part of
an iron-rich basalt flow suggeststhat the hornll
14HO
blende may have grown as a quench phase. and
I
was probably metastable. The differentiation of
the flows can be modelled by crystal fractionaA L L A s s e r v r g u a ] e IsN C L U D E
Q U A R T Ztion in which hornblende plays a key role. In the
A N DM A G N E T I T E
iron-rich flow, crystal settling of hornblende
(40Vo by weight added to the parent composiFlC. 8 The invariant
point quartz-magnetitetion) produces the most extreme cumulate comepidote-actinolite-pumpellyite-prehnite-chlorite
at
position. We feel that the iron-rich naturb of
constant 1, below a threshold value of pco, which
these basalts, which are similar to the ones in
would add calciteto theobserved assemblage.
Garrison Township to the north, probably re$lid circle indicates the observed assemblage.
flects the original composition rather than fracQ-quzxtz, M-magnetite, B.epidote, A-actinolite,
P-pumpellyite, Pr-prehnite. Cl-chlorite
tionation of a more "primitive" magma type.
,1
I
I,IETAVOLCANIC
ROCKS
FROM
THACKERAY
AcrNowleocEMENTs
This research was supported financially by
Queen's Research Awards 279-AL9-OI and 273019-31 and National Research Council of Canada Operating Grant 48709. We are indebted
to Larry Jensen of the Ontario Division of
Mines for showing us the area, and for useful
discussion in the field, particularly concerning
the 'oteverse-diabasic"texture. B. E. Gorman
provided able field assistanceand critical discussi'on.Critical reading by P. L. Roeder, D. M.
Carmichael and L. D. Ayres resulted in improvementsin the manuscript.
REFERENcES
Coorans, D. S., HonooysKr, R. J., & Nevron, R. S.
(1970) : Occurrence of prehnite-pumpellyite facies
metamorphism in northern Maine, Am. l. Sci.,
268. 142.
ENGn, A. E. J., ENcrr", C. G., & HevrNs, R. G.
(1965): Chemical characteristics of ocean basalts
and the upper mantle, Geol. Soc. Am., Bull.,76,
71,9.
FRArrA, M., & Sne.w, D. M. (1974): 'Residence
contamination' of K, Rb, Li, and Tl in diabase
dikes, Can. l. Earth Sci." 11" 422.
Gooowrx, A. M. (1972): Plate Tectonics arul
Evolution of Frecambrian Crust, in Implications
of Cantinental Dri,It to the Earth Sciences, ttol. 2,
edited by D. H. Tarling and S. K. Runcorn,
Academic Press, London and New York.
TOWNSHIP,
ONTARIO
519
JrNsrN, L. S. (1969): Bisley Township, District of
Timiskaming, Ontario Div. Mines Prel, Map
P520.
(l97la):
Clifford Township, District
of Timiskaming, Ontario Div. Mines, Prel. Map
P692.
(l97lb): Elliott Township, District of
Cochrane, Ontario Div. Mines, Prel. Map P703.
(1973): Thackeray township, District
of Cochrane, Ontario Div. Ml;nes, Prel MapP843.
Jorrv, W. T. (1974): Regional metamorphic zonation as an aid in study of the geologic history of
an ancient archean terrain: Abitibi Region, Ontario-Quebec (abs.), Geol. Assoc. Can.l Mineral.
Assoc. Can,, Annual Meeting and Abstracts, 46.
Jor.rv, W. T., & Svrrur, R. E. (1972): Degradation
and metamorphic differentiation of the Keweenawan tholeiitic Lavas of northern Michigan,
U.S.A., Jour. Petrol., 13, 273.
Ltou, J. G. (1971): Synthesis and stability relations
of prehnite, Ca:,AlzSiuOro(OH)2,
Ant, Mineral., 56,
507.
PEARcE,T. H. (1968): A contribution to the theory
of variation diagrams, Contr. Mineral. Petrol., 19,
r42.
(1970): Chemical variations in the
Palisades sill, Jour. Petrol. Ll, 15,
Slvrrru, R. E. (1967): Segregation vesicles in basaltic lava, Am. Jour. Sci., 265, 696.
(1968): Redistribution of major elements in the alteration of some basic lavas durins
burial metamorphism, Jour. Petrol., 9, l9l.
Manuscript recei,vedJuly 1974, emended September
1974