Archaean Deformation Patterns in Southern Africa

Author(s): M. P. Coward
Source: Philosophical Transactions of the Royal Society of London. Series A, Mathematical and
Physical Sciences, Vol. 283, No. 1312, A Discussion on Natural Strain and Geological Structure
(Oct. 12, 1976), pp. 313-331
Published by: The Royal Society
Stable URL: http://www.jstor.org/stable/74647 .
Accessed: 08/05/2014 06:31

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Philosophical
Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences.

http://www.jstor.org

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
Phil. Trans.R. Soc. Lond.A. 283, 313-331 (1976) [ 313 ]
Printedin GreatBritain

Archaeandeformationpatternsin southernAfrica

BY M. P. COWARD
Departmentof Earth Sciences, Universityof Leeds

Strain measurements have been made to help quantify the intensity of deformation
and amount of displacement across Archaean greenstone belts in Rhodesia and
Botswana and across the gneisses of the Limpopo mobile belt. The area has been
divided into three domains based on the orientation of the finite strain fabric and
the orientation of the maximum extension direction in associated shear zones. The
domains are considered to have different movement patterns and to be similar
to small orogenic belts.
Early deformation within the greenstone belts accompanied the intrusion of the
diaipric granites, but there was also bulk translation and rotation of greenstone belt
and gneiss leading to imbrication of the stratigraphic pile and the formation of large
nappes of overturned rock. This was followed by regional phases of deformation which
affected all the greenstone belts and the gneisses of the Limpopo belt. Detailed strain
measurements show a variation in amount of shortening during this phase, from
under 30 % across the Shabani-Bellingwe belt in central Rhodesia, to over 60 % across
the Tati and Matsitama belts in northern Botswana. Many local variations in
intensity of deformation occur within large ductile shear zones and deviations from
plane strain may be partly due to such rotational deformation. The regional deforma-
tion pattern suggests that there was movement of the Rhodesian craton approximately
200 km to the southwest relative to the gneisses of the Limpopo belt, producing a
dominantly flattening deformation in the southwest of Rhodesia, but dominantly
simple shear with a nearly horizontal sinistral movement, in the southeast.

INTRODUCTION

In the Archaean rocks of southern Africa, measurements of natural strain have previously
been confined to small areas or to individual problems. Ramsay (I963) reported two strain
measurements from the Fig Tree and Moodies formations of the Barberton greenstone belt in
the eastern Transvaal and Gay (I969) and Anhaeusser (I966) give further strain measurements
from the Fig Tree formation in this belt. Wood (1973) gives results of ten strain measurements
from the conglomerates of the Umvuma greenstone belt in central Rhodesia.
In 1970 the Research Institute of African Geology at Leeds University began a project to
examine the geological structure and quantify the strain across the Limpopo orogenic belt and
the greenstone belts and granites in the adjacent Rhodesian and Kaapvaal cratons (figure 1).
Some results have been given elsewhere, in Coward & James (I974), Coward, James & Wright
(in the Press), James (I975), Graham (I974) and Wakefield (I974). The aim of this paper is
to summarize these results and discuss the geological structure of the Archaean rocks of
southern Rhodesia and northeast Botswana in the light of these strain measurements.
From geochemical evidence, Anhaeusser et al. (1969) consider that the greenstone belts of
Rhodesia and northeastern Botswana (figure 1) were originally deposited on ocean floor
without any basement granite crust. However, Oldham (1970) and Bickle et al. (1975) report
a clear unconformity of greenstone belt sediments resting on a basement gneissic granite,
and Coward, James & Wright (in the Press) outline several areas of this gneissic basement.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
314 M. P. COWARD

FIGURE 1. Simplified location map of part of southern Africa showing Rhodesian and Kaapvaal cratons, the
greenstone belts and the position of the granulites at the northern edge of the Limpopo Belt. The position
of the northern margin of the Limpopo belt is diffuse and depends upon the bias of the author. The line
of this figure is after Mason (I973). The southern boundary is equally diffuse. Van Breeman & Dodson
(I972) place it north of the granulites in the Transvaal; Mason (1973) places it south of the granulites.
The 'Cover' is the Witwatersrand sediments and younger rocks (less than 2600 Ma).

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 315

Hickman (1974a) and Hawksworth et al. (1975) give ages of approximately 3600 Ma for these
old gneisses.
The greenstone belts show a lithostratigraphic sequence from ultrabasic and basic volcanics
to acid volcanics and sediments (Wilson I973). Coward, Lintern & Wright (in the Press)
consider that all the greenstone belts in southwest Rhodesia were once part of a large basin in
which there were lateral facies changes. In the east, shelf sediments, intercalated with basic
volcanics, rest on basement gneiss while in the central part there are dominantly ultrabasic and
basic rocks. In the west there is a wide zone of thick shelf sediments which crop out in the
Matsitama belt in northeast Botswana and consist of current-bedded sandstones, limestones
and some basic volcanics. These are correlated with similar sediments in the Limpopo belt
at Shashi and Madinare in eastern Botswana. Thus the metasediments of the Limpopo
belt, which are normally termed the Messina Formation (Sohnge et al. I949) from their type
occurrence at Messina in the northern Transvaal, have been through the same deformation
sequence as the main Rhodesian greenstone belts and appear to be of similar age but different
sedimentary facies.
The greenstone belts and the sediments of the Messina formation have been intruded by
several phases of diapiric granite, tonalite and diorite. The earliest, granites are deformed to-
gether with the greenstone belts but the later granites form large irregular-shaped batholiths.
Hawksworth et al. (I975) obtained ages from four of the greenstone belts and three of the
granites from central Rhodesia of between 2800 and 2500 Ma, and similarly Hickman (1974b)
has dated three of the large batholiths at 2600 Ma.
In the northeastern part of the Limpopo belt there is an area of granulite facies rocks
(figure 1) (Cox et al. 1965; Mason 1973). Satellites of the Great Dyke, which intrude the granu-
lites, have been dated at 2580 Ma, giving a minimum age for this granulite facies metamor-
phism (Robertson & van Breeman 1970). Deformation in the central part of the Limpopo
belt may post-date the Great Dyke; gneisses in eastern Botswana, give a thin slice, whole rock
isochron of 2100 Ma and Hickman & Wakefield (I975) consider that this indicates the age of
the latest phase of ductile deformation in this zone.

THE FINITE STRAIN FABRIC AND STRAIN MEASUREMENT

The granites and greenstones of Rhodesia and the gneisses and metasediments of the Limpopo
belt show a penetrative finite strain fabric which can be traced from Fort Victoria south to
Messina and from Shabani west to Matsitama. In the greenstone belts this fabric is shown by
the preferred orientation of hornblendes, chlorites and micas, by pressure solution stripes and
by the shapes of deformed objects (Coward & James 1974). In the granites and gneisses it is
shown by the preferred orientation of minerals and the shape and orientation of deformed
grains and inclusions. Measurements of the shape of the deformation ellipsoid were made from
ellipsoidal or sub-ellipsoidal objects using methods described and Ramsay (1967), Dunnet
(i969), Dunnet & Siddans (197i) and Roberts & Siddans (1971). Some quartz-rich grits and
quartz grains within deformed granites were measured by the centre-to-centre method
described by Ramsay (i967, pp. 195-197), that is by measuring the distance between centres
of adjacent grains, this distance being proportional to the longitudinal strain in that direction.
The shape of the strain ellipsoid was also measured using the method described by Talbot
(1 970) which involves the determination of the shape of the surface of no finite elongation from a

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
316 M. P. COWARD

stereographic plot of poles to folded and non-folded or boudinaged veins, the boundary between
the two fields being the trace of the normal to the surface of no finite elongation. A simple
modification of the Talbot method was used in which the orientation of the surface of no finite
elongation was measured directly on a surface, preferably parallel to a principal plane, by
measuring the pitch on this surface of the line which separates those parts of the veins which
are folded from parts which are boudinaged. Where there were insufficient veins for such de-
terminations, measurements were made of the shortening along folded veins by using the
method described by Hudleston (I973). This involves measuring the amount of post-buckle
flattening and the shortening due to fold amplification and estimating the amount of shortening
prior to fold amplification from the characteristic fold wavelength/thickness ratio. There may
be considerable errors in such estimates; Cobbold (this volume) has shown that natural folds
need not develop with a characteristic wavelength and hence estimates of the pre-buckle
shortening may be little short of guess-work.
The strain parameters used in this paper are the ratios of the three principal axes, X > Y > Z,
the percentage shortening or elongation in these directions, assuming constant volume, the
natural logarithmic octahedral strain, Es (a function of the octahedral unit shear yo) and the
Lodes unit V where
Yo = 4 (In xly + In ylz + In x/z) ,

Es = (-3) o

and V = (ln y/z - ln x/y)/ln x/z

(Hossack 968; Gay 1969). Values of Vrange from + 1 for uniaxial oblate ellipsoids to - 1 for
uniaxial prolate ellipsoids.

SUBDIVISION OF THE REGION INTO STRUCTURAL DOMAINS

The fabric in the granites, greenstones and gneisses is a finite fabric, a result of several de-
formation stages, not necessarily synchronous throughout the whole area. The orientation of
the finite fabric varies in orientation as does the mineral lineation which is normally parallel
to the maximum extension direction of the deformation ellipsoid. The area has been divided
into different domains based on the orientation of this fabric and on the orientation of the
maximum extension direction in associated shear zones. The domains are considered to have
different movement patterns similar to small orogenic belts, and are numbered 1 to 3, domain 3
being the youngest.
Domain 1 is characterized by a steep foliation and a down-dip lineation in the west, but in
the south and east, the foliation curves form a northwest to a northeast trend, while the mineral
lineation and maximum extension direction plunges northeast or southwest throughout
(figure 2). There are localized zones of more intense deformation, major shear zones, which
cut across the regional foliation. In the southern part of the domain, the shear zones have
a northeast trend, a sinistral sense of movement and a nearly horizontal maximum extension
direction. In the north and west of the domain, the shear zones strike north-south and have
a dextral sense of movement but the lineation plunges in the same northeast-southwest plane
(figure 2). Coward, James & Wright (in the Press) consider this arcuation of foliation and change
in sense of shear zones to indicate movement of the Rhodesian craton to the southwest relative

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 317

to South Africa, producing shortening across the foliation in the southwest part of the domain
and some shortening but also considerable simple shear with near-horizontal movement in
the southeast of the domain (figure 3). Much of this paper will be concerned with the state of
strain within this domain.

FIGURE 2. Map showing the trend of the foliation and lineation within the three structural domains. Domain 1 a
is subdivided from domain 1 on the presence of later structures which fold the cleavage. Cover is shown
clear.

The northern margin of domain 2 is marked by a major shear zone with a down-dip lineation.
Within this domain, the foliation has a similar trend to that in the adjacent parts of domain 1,
however the fold hinges, the lineation and the maximum extension direction are near-normal
to those in domain 1 and plunge to the south-southeast (figure 2). The rocks are granulite-facies
gneisses and metasediments which are deformed by several major shear zones, each with the
same strike of foliation, down-dip lineation and sense of movement. James (1975) has described
the deformation of this domain in terms of a regional but heterogeneous simple shear deforma-
tion, the granulites being uplifted and thrust from the south-southeast.
The northern margin of domain 3 crosses the boundary of domains 1 and 2 (figure 2). In the
east, the junction between domains 3 and 2 is a zone of gently dipping mylonites, while in the
west the junction between domains 3 and 1 is a steep shear zone (Coward et al. I973). In the
east where this shear zone cuts relatively undeformed granites and granulites, it forms a new,
near-horizontal fabric, but elsewhere, the earlier fabrics are rotated to produce an intense,
generally mylonitic, finite strain fabric. On the mylonite foliation surfaces there is a well
developed lineation shown by elongate quartz leaves and feldspar augen which presumably
marks the movement direction in the shear zone. From this lineation and the curvature of the
foliation into shear zones, the sense of movement is inferred to be dextral, southern and upper
gneisses moving west, parallel to the strike of the shear zone (figure 3). In the east, the
mylonitic gneisses grade upwards and southwards into less intensely sheared but still
dominantly flat-lying gneisses which are interbanded with sediments of the Messina formation.
These flat-lying rocks are folded and locally intensely flattened into a series of upright peri-
clinal folds with axes trending north-northeast. These folds cannot be recognized in domain 2

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
 ,.

318 M. P. COWARD

or in the southeastern part of domain 1 and hence the marginal shear zone has acted as a plane
of decollement between folded rocks to the south and the granulites, granites and greenstones to
the north. Strain measurements made in the mylonites with the centre-to-centre method on the
shape of deformed quartz grains gave an estimate of up to 50 km displacement across this zone
in southern Rhodesia (Coward, James & Wright, in the Press; James 1975). As much of the
deformation in the more cataclastically deformed gneisses presumably took place by grain
boundary sliding or shear along discrete planes, strain measurements made by any conventional
method are likely to be much too low, and as much of the flattening in the gneisses with domain 3
to the south may also have been accompanied by simple shear, this figure of 50 km displacement
is probably a considerable underestimate.

(a (b) i

~
2/./

_ __ .,. _j
. ..... ..//
I~
Iu
. /.. i
_2 _ ___'\ ' -

the deformation of the three domains.

DEFORMATION HISTORY OF DOMAIN 1

The deformation state of the rocks in domain 1 is a result of several deformation events. Though
the main fabric in this domain is a finite fabric and not necessarily synchronous throughout the
area, it can be used as a form of time marker for the deformation sequence. The deformation
history can be separated into:
1. A pre-cleavage regional deformation, prior to the intrusion of the diapiric granites.
2. Deformation associated with granite intrusion.
3. Regional deformation producing the main fabric.
4. Late phases of deformation which fold and whose fabrics cross-cut the regional fabric.

1. The pre-cleavageregionaldeformation
The rocks of the Tati greenstone belt in northeastern Botswana dip to the southwest but are
overturned and young to the northeast (Mason I968; Coward & James 1974). Litherland &
Key (I974) have shown that the Tati greenstone belt can be joined to smaller fragments of green-
stone belt material to the north (figure 4), to make an extensive sheet of overturned north-
eastward facing rocks. This overturning pre-dated the intrusion of the diapiric granites and the

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 319

development of the main cleavage (figure 5). Similar overturned rocks occur in the core of the
Matsitama belt, west of Tati (figure 4).
In southern Rhodesia, the Antelope and Lower Gwanda greenstone belts young upwards
to a layer of gneiss which shows several phases of complex early deformation and which may
well be allochthonous basement gneiss (Coward, James & Wright, in the Press). In southern
Rhodesia, the Selukwe greenstone belt was overturned and thrust into its present position

FIGURE4. (a) Map of the southwest part of domain 1 showing the autochthonous and allochthonous greenr
stone belts and gneisses Greenstone belts: M, Matsitama; V, Vuma; T, Tati; B, Bulawayo; G, Gwanda;
LG, Lower Gwanda; SH, Shabani.
(b) Schematic section through the greenstone belts between Matsitama and Bulawayo, before the
intrusion of the diapiric granites. Arrows indicate younging direction.

before the intrusion of the diapiric granites (Stowe 1974). Thus much of the greenstone belt
material and the gneiss in the south-west part of this domain is allochthonous and much of this
early deformation took place by large scale bulk translation and rotation. The Selukwe green-
stone belt may have been transported some considerable distance, the nearest comparable auto-
chthonous ultrabasic and basic rocks are more than 50 km to the south or southwest, south of
Belingwe and at Filabusi.
The greenstone belts at Bulawayo, Shabani and Fort Victoria appear autochthonous and
that at Shabani rests directly on older basement, but these autochthonous greenstones also

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
320 M. P. COWARD

suffered folding before being intruded by granites and were then deformed by the cleavage-
producing episode (figure 4). Although this folding was locally tight, no penetrative fabric was
produced.
No thrust planes are seen except in the Matsitama belt where thejunction between uninverted
and inverted beds is marked by a zone of intense stylolitization.
granite - basic acid
volcanics volcanics
porphyritic
---- n r-
centre - ... -

foliated granite with // /./: . i

mega-xenoliths of/ // /It // /
g
granite
Messina type Z !-/ //v/ *//11/ with
sedime1nt .1/1 f .1/7 / / I oshears
/}: j
f l>1zones
/ only weakly
/deformed
intenset
A xenolith of greenstone deformation A'

ultrabasics
basics
and acid
granite ironstones sediments volcancs

strip of '
basic rock - as
xenolith in granite

B F3 Synform)

FIGURE 5. Two cross sections through the Tati greenstone belt.

Section lines are shown in figure 4.

2. Deformationassociatedwith the intrusionof granite

Anhaeusser et al. (I969), Anhaeusser (I973) and Mason (I973) considered that much of the
deformation in Archaean greenstone belts was caused by the forceful intrusion of diapiric
granites. However the granites in Rhodesia and northeast Botswana clearly post-date the over-
turning of the Tati and Vumba greenstone belts (Litherland & Key I974; Coward & James
I974) and at Bulawayo and Fort Victoria the granites post-date the folds in authochthonous
greenstones.
Macgregor (195I) introduced the term 'gregarious batholith' for the large masses of
'granite' which surround the greenstones, and Fyfe (I973) suggested that the size of these
batholiths reflected the size of the convection cells in the Archaean upper mantle. He suggested
that the batholiths represent granite domes which grew by addition of new material at the
centre, while the edge expanded outwards as a balloon.
However, most batholiths cannot be considered as simple 'balloons'. They consist of several
ages of granite and the latest granites intrude not only the earlier granites at the centres of the
batholiths, but also greenstones at the edges. Much of the 'granite' within the batholiths is
basement gneiss, and this maintains its early tectonic trend throughout large areas and even has
the same trend in adjacent batholiths (Coward, James & Wright, in the Press). There seems to

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 321

have been very little reorientation of this gneissic material during the formation of the batho-
lith, and most 'gregarious batholiths' should not be considered as batholiths at all.
The intrusive granites occur in two main forms:
(i) As distinct bodies, circular or elliptical in plan. These bodies vary from 1 to over
30 km in diameter. They are probably sections through diapiric 'bubbles' as described by
Fyfe (i973).
(ii) As large, often irregular bodies, whose edge may be partly fault-controlled. The Chibi
batholith south of Shabani forms one such body, over 30 km wide and over 140 km long.
The granites are of various ages. Many carry a foliation throughout, not just at the margin,
and this foliation appears continuous with the tectonic foliation in the adjacent greenstones
(Coward & James I974). Many of the diapirs are elliptical in plan, their long axes being
parallel to the foliation trend. Thus many of the granites have suffered a regional deformation
along with adjacent greenstones. Other granites clearly post-date all the deformation; they
cross-cut the structures and have irregular unfoliated margins.
Faulting and stoping were important intrusion mechanisms. Many of the early granites,
such as those north of the Tati greenstone belt (Litherland I973) contain large masses,
'mega-xenoliths', of greenstone-belt material. Some of the early gneissic material was reacti-
vated and uplifted by faulting. West of the Fort Victoria greenstone belt, the Chibi 'batholith'
which consists of late post-tectonic granite and also early, possibly pre-greenstone gneiss, cuts
across all the structures in the Fort Victoria belt, but produces no new visible fabric in granite,
gneiss or greenstone. The contact must be an arcuate fault or a series of faults.

higher than that in the adjacent greenstone belts. The deformation shown by these xenoliths
consists of two components, the regional cleavage producing deformation and an earlier defor-
mation due to the intrusion of the igneous body. This is shown by the skewed distribution of
points on the plot of strain ration (Rr) against s, the pitch of the long axis of the strain ellipse.
By using the programme 'STRANE' (Dunnet & Siddans I97i) and assuming that the cleavage
represents the principal plane of the later deformation, values have been assigned to the two
components of the finite strain in a tonalite southwest of Gwanda (figure 9). Though these
tonalites show internal deformation, it is difficult to know how much of this deformation was
transmitted to the adjacent country rock.
Some diapirs may have been formed by forceful intrusion; a rim syncline surrounds a granite
which intrudes the Mont d'Or area of Selukwe, though this deformation may have been
accentuated by later deformation. Similarly the greenstones in the southern part of the Tati
belt were deformed to a distance of 1 km away from the contact of a large tonalite body before
the main cleavage-producing deformation (Coward & James 1974). At distances greater than
1 or 2 km from the contacts, there seems to have been very little deformation due to granite
intrusion.
Within the Limpopo belt, the cleavage-producingdeformation was preceded by the segrega-
tion of quartzo-feldspathicmaterial, granite and pegmatite and the intrusion of large sheets of
granite. Coward,James & Wright (in the Press) suggest that many of these migmatites may be
the deeper level equivalents of the diapiric granites seen in central Rhodesia. However, recent
isotope work on the Rhodesian granites (Hawksworthet al. 1975, Moorbath, in the Press)shows
that much of the Rhodesian granite is probably mantle-derived and cannot have been formed
by partial melting of much older basement gneiss.

21 Vol. 283. A.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
322 M. P. COWARD

In the area between Bulawayo, Fort Victoria and the northern edge of the granulites, green-
stone belts make up 28 % of the rock, older basement gneiss 18 %, early intrusive granite 25 %
and younger unfoliated intrusive granite 29 %. Thus if the implications of Hawksworth et al.
(1975) are correct, over 50 % of the rock now exposed at the surface in southern Rhodesia was
new intrusive material added to the crust and not derived from partial melting of basement
gneiss. This raises a severe space problem which has not been solved. Stoping and fault-
controlled intrusion must have been the most important mechanisms for granite emplacement,
but a considerable amount of material must have either sunk within the magma or must have
been uplifted by the incoming granite. A large part of Rhodesia must have been underplated
and possibly uplifted during this period and the tectonics may have been similar to those at
the western margin of South America at the present day.

0\6)"'KN (a) (b)

g LOWERt-42A Ti a SH A I

<d ^/ \
VICTORIA
X, f ' :
:^ .--_
belt were t AWAYOunt ofdata, and not
GWAre represented in figure 6a.
Wright, in the Press) an te -49z2
-y\\ v-^vx \TAI -ly V
-60z
.SHABANU \-

ll \ -
\i 1 I \
0.5 1.0 1.5
In(Y/Z)
FIGURE 6. (a) Mean strains from five greenstone belts in domain 1, on logarithmic deformation plot (data from
Coward, James & Wright, in the Press). Lines of equal value of V (Lode's unit), E. (solid lines) and
percentage shortening in the z direction (dashed lines) are shown.
(b) Map showing the mean strain expressed in terms of percentage shortening in the z direction and
percentage shortening or elongation in the y direction for the greenstone belts. Figures from the Matsitama
belt were taken from a smaller amount of data, and are not represented in figure 6a.

3. Regional deformationproducingthe mainfabric

Few major structures were produced during this phase of deformation, but the earlier struc-
tures and the diapiric granites were deformed to produce the regional finite fabric. Strain
measurements have been made on greenstone belts (Coward & James 1974; Coward, James &
Wright, in the Press), and the results are summarized in figure 6. The intensity of deformation
varies from over 60 % shortening across the Tati greenstone belt to under 30 % across the
greenstone belt at Shabani. East of Shabani the deformation is negligible; the rocks are tilted
but show no cleavage, the unconformity of sediments underlying greenstones resting on base-
ment granite is preserved intact and sedimentary structures such as ripple marks and stroma-
tolites are tilted but otherwise undisturbed.
Measurements of the amount of shortening along buckled veins are given in figure 7 for
14 localities in the Tati greenstone belt, 7 localities in the gneisses south of Gwanda and
5 localities in the gneisses southeast of the Lower Gwanda-Antelope belt. These values are

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 323

TATI F2 GNEISS NORTHOF LOWER GWANDA

4 - 14 LOCATIONS 7 LOCATIONS

I L
L I
I II

TATI F3 GNEISS SOUTH OF LOWER GWANDA

4 _ 5 LOCATIONS 5 LOCATIONS

2-

_
I I I
0 40 80 0 40 80
FIGURE7. Histograms showing the amount of shortening along buckled veins.

II
II t
II
2
I

E, /',.
I 11
I
tI
0
*
/
/ I

1
6\

P
-t-1

I
1i I \
*s I '
\\

Bv<o
>V>O
V >0.5

FIGURE 8. Map of the Tati greenstone belt showing the main shear zones, the variation in oblateness of the
strain ellipsoid (variation in Lode's unit, V) and a strain profile through part of the northwest arm of the
belt.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
 324 M. P. COWARD

>
GWANDABELT -k foliated granite (b)
\S^~ ~\ '~
, ---_~ ^~R, tectonic strain
(a) R,S1.8 Rg intrusive strain
Rg 2.85 *Indeterminate -^ younging direction 1.5
.5
OWERGWA-DA
8.0 +

Rg 4.02 +._ _
1
~R, 0+ +.5 __
'
sxRg 4.18 + + younger non-foliated
+ + + . granite

A NTELOPE- -'"~
.
""-~ + "-- +
LOWER GWANDA . +
BELT s <t

A'--..

%.'- mugs,early thrust plane i=. =-? ' _

00.52.
A B
(C pld) Claaei hw ynro ie0.1km

N i Ego^ \0.2 --

measurements
bythe solidcircles.(Parly
shown Alu

mylonite
FIGRE 99.(a) Map
FIGURE((b) Mapofofthe northern
the northe ofGwanda-Antelope
part of the Lower
the Lower greenstone belt showingbelt. the how the strain
how strain
eastern
part Gwanda-Antelop showing be
ratios measured
ratios deformed
from deformed
measured from xenolit
xenoliths on horizontal
hs on horizontal older foliated
in thesurfaces in the granite
surfaces older foliated can
granite can be
separated into two components, a tectonic strain (R,J) and an earlier strain due to the granite intrusion
(R,,). The line indicates the orientation of the principal axis of the R. strain ellipse. Insufficient outcrop
prohibited measurements in three dimensions. In the eastern part of the map, the basement gneiss is shown
st'ippled. Cleavage is
stippled. Cleavage is shown
shown by
by narrow lines.
narrow lines.
(b) Strain profile through the eastern arm of the Lower Gwanda belt. Location of profile section shown
in figure 9 a.
(c) Map of the Bulawayo greenstone belt showing the zone of more intense deformation at the front of
a shear zone. Small syenitic granites in the Bulawayan greenstones are shown (g). Locations of strain
measurements shown by the solid circles. (Partly after Amm I940.)

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 325

a little higher than the values of shortening calculated from measurements of deformation
ellipsoids assuming no volume change. This discrepancy may be due to volume change but is
more likely to be due to errors involved in measuring the buckled veins and estimating the
amount of shortening before fold amplification using a characteristic wavelength/thickness
ratio.
At Tati, Lower Gwanda and the southern part of the Belingwe belt, measurements of the
deviatoric strain made from deformed objects were combined with measurements of the surface
of no finite elongation made from the orientation of buckled and boudinaged veins. These
measurements enabled the volume change to be calculated (Barr & Coward I974). Measure-
ments from the basic rocks at Lower Gwanda and Belingwe give a volume loss of under 5 %,
locally up to 12 %.

FIGURE 10. Map of the granulites in the southern part of domain 1 and the western part of domain 2. Lithologi-
cal layering is shown by heavy lines, cleavage by narrow lines. The circles are localities of strain measure-
ments. T, Tod's Hotel. Eastern portion of map after James (I974).

Apart from the regional variation in strain intensity shown in figure 7, there are local
variations within greenstone belts. In the northwest arm of the Tati belt (figure 8), a shear zone
can be seen from the curvature of the foliation and from the localized more intense strain.
Similar variations occur in the Lower Gwanda belt (figure 9 a, b). There is intense deformation
along the southern margin of the Bulawayo greenstone belt, with the production of a mylonite
fabric. This shear zone dies out along strike to the west and the deformation is taken up by
21-2

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
326 M. P. COWARD

extra flattening within the Bulawayan greenstones (figure 9c). This extra flattening can be
measured from the shape of deformed agglomerate fragments and also from the shape of small
deformed granite diapirs which are elliptical in plan in the zone of more intense deformation
(figure 9c).
There are also variations in intensity of deformation in the granulites in the southern part of
the domain (figure 10). Much of this rock is high-grade, sometimes charnockitic granite with
very little tectonite fabric and much of the granite contains undeformed xenoliths of basic or
earlier gneissic material. These relatively undeformed granites are cut by steeply dipping shear

ii
2I I1
I ! iI f
-I 9/ . t O I I L
* ! I ll -

I
E _ / / +1

-16
0 1 ln(YIZ) 2 P Q
FIGURE 11. (a) Strains measured from the granulites and adjacent gneisses on logarithmic deformation plot.
(b) Strain profile through the southern part of domain 1, section line on figure 10. The horizontal bars
represent areas of low-intensity deformation as seen from the paucity of tectonic fabric and relatively
undeform2dxenoliths
undeformed in the
xenoliths in charnockiticgranites.
the charnockitic granites. 20
The horizontal
The horizontalarrowed line is
arrowedline is 20 km.
km.

zones with an intense tectonite fabric. The strain has been measured from the shape of the
xenoliths and from the surface of no finite elongation from folded and boudinaged veins, and
estimated from measurements of the shape of quartz leaves; esimte shapeeaurmetso
thenro te hae fquartz
of the original quartz
grains changes and becomes more ellipsoidal within the shear zones. The results are plotted in
figure 11. In the eastern part of the granulite outcrop there is less granite; the rocks are more
gneissic and are folded with low-intensity deformation in the hinges but intense deformation in
the limbs (figure 10). Shearing has been concentrated in the limbs of these folds, producing a
'mega-lithon' type of structure like that seen on a small scale in crenulation cleavage.
In the greenstone belt at Lower Gwanda, the finite deformation is close to that of plane strain,
but in the other belts the ellipsoids are more oblate, while in the granulites in the south they are
more prolate. This deviation from plane strain may be partly due to rotational deformation.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 327

In the Tati belt, increments of simple shear have been added to the regional deformation and
the movement direction in the shear zone is normal to the maximum extension direction of the
regional deformation, but nearly parallel to the intermediate extension direction. The result
is an increase in oblateness with increase in intensity of simple shear (Coward & James I974).
There do not appear to be any significant regional variations in elongation in the inter-
mediate extension direction which might indicate that the arcuate foliation was due to later
bending. There is no shortening in this intermediate extension direction on the inner part of
the arc, nor any major increase in extension in this direction in the outer part of the arc. The
least deformed rocks occur in the inner part of the arc at Shabani, while the most intensely
deformed rocks are in the southwest of the domain. Thus the strain measurements support the
contention that the arcuation is a result of an undeformed 'protocraton' in central Rhodesia
moving southwest relative to adjacent areas, producing extensive flattening in the southwest
of the domain but dominantly simple shear with a sinistral sense of movement in the southeast
(figure 3).
Assuming that the anomalously intense deformation in the shear zones can be ignored, and
assuming irrotational deformation between the shear zones, a mean shortening of 49 % has
been computed for the section between Matsitama and Shabani. This computation assumes
that the deformation in the granites and gneisses is similar to that in the adjacent greenstone
belts, as it is only in these belts that abundant reliable strain markers have been found. Support
for this assumption is given by the agreement between the amount of shortening shown by
buckled veins in gneisses adjacent to the Lower Gwanda belt and the amount of shortening
shown by deformed objects within the belt (figures 6 and 7).
If all the above assumptions are correct, there must have been movement of the Rhodesian
craton some 200 km to the southwest relative to the gneisses of the Limpopo belt. However
much more of the deformation may be rotational than has been assumed and this figure
probably represents a maximum for the shortening during this phase.

4. Strainsdueto latephasesof deformation

Late phases of deformation are important in the west of domain 1 where the main foliation
is often crenulated and locally tightly folded (figure 5). The amount of shortening measured
along veins folded during this phase, from five locations at Tati, are shown in figure 8. Late
folds which fold the cleavage are important in the gneisses south of Tati, where they become the
dominant structures in the northwest part of the Limpopo belt. They are also important in
the Matsitama belt, where strain measurements indicate intense deformation, with X-Z ratios
of over 30:1, partly as a result of this late deformation.
In central Rhodesia, late folds are important in the Selukwe and Shabani areas. At Selukwe,
the cleavage-producing deformation is intense in the south, but north of Selukwe township it is
very weak and conglomerates in the Selukwe greenstone belt are undeformed. The later folding
took place on the north-northeast trending axial planes and deformation is accentuated in shear
zones with a similar trend. There are two major shear zones. The Wanderer shear zone in the
east has a down-dip lineation. This intense shearing deformed conglomerate pebbles into
ellipsoids with the same down-dip maximum extension direction and where the shear zone cuts
the ultrabasic rocks south-east of Selukwe, chrome-bearing ore bodies are prolate with the same
down-dip extension (cf. Cotterill 1969). The shear zone wraps around the granites producing
21-3

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
328 M. P. COWARD

Q 5 km....... ............. ...........

....
11 ..... ii. ?.......
1-. f -l- .-*. .
...........-

a i
'. \ \ :.......... \II
...;:r:;:;::::::::;::::::::::~~~~~~~~~~:?.,?.?.?........
:!:i::~~~~~~~~~~~~~~~~~~~..:
........... - .
.
..... ........::?:?::
??::
*?-;:1
.?~3:f~:j
----------
!
.....
................ .............:
..................
.....?~:: ...?'=
'' j
..... 11---
............ ....
..... .......
....
iii~~~~~~~~~~~~~~~~~~~..........
..........
r OL'.B 1-..
+~~~~~~~~~~~~~~~~~.........
-`r;-I-:*:?:?[?:l:?:?;?:?:?:?;?.?r.'.'.'
,
5 km .. .............
................
........
............
..............
.......
.......
.- O 1 .::........
;f;?f-
jinv~~~~~~~~~~~i
:..+.
" +..+ + t ....-t........
..............
f -f? G
-f~~~~~~~~~~~..........
?i---ct;f $- + C

..............
.......
..............-

+ ++ + +

//' ? + +
I~ ;Z +
~ y+
1('Ii

~+ ++ +I + +6?

T7-
N
ii 'i;'

I
k
..N-l

.// cleavage trace

i i lineation
deformedconglomerate
ET3 Great Dyke
El Adamellite
I\ Montd'Orgranite
Gneissic
LIIItSedimentsgranite
LIII..
M.
&basicvolcanics
Ultrabasic rock &chromite

FIGURE 12. Map of the Selukwe greenstone belt, partly after Stowe (1968).

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 329

the most intense deformation at the eastern and western margins of the granites, but leaving
areas of low-intensity deformation, pressure shadows, at the northern and southern margins
(figure 13).
West of Selukwe, the Surprise shear zone forms a zone of intensely deformed rock more
than 1 km wide (figures 12 and 13), which cuts across the granite bodies. The maximum exten-
sion direction is nearly horizontal and the sense of movement is dextral. Small adamellitic
bodies have been mapped by Stowe (1968) west of Selukwe. Away from the shear zone, these
bodies are irregular in shape an(d orientation, but they become elliptical in plan towards the
shear zone, and show a strain ratio of up to 10: 1, similar to that given by the shape of deformed
quartz grains in the granites and gneisses within the shear zone.

FIGURE 13. Map showing the two principal shear zones. S, Selukwe.

South of Selukwe, all the structures and fabrics described above are deformed by a ductile
shear zone trending east-west, with a sinistral displacement of some 8 to 10 km (Coward, James
& Wright, in the Press). North of Selukwe, another east-west trending zone deforms the Um-
vuma greenstone belt east of the Great Dyke. Wood (I973) recorded intense strains wth 75 %
shortening in the z direction and 400 % extension in the x direction from deformed conglo-
merates within this belt.
The Great Dyke and its satellites cut these east-west shear zones but lie parallel to the north-
northeast trending structures; the satellites of the Great Dyke mark the eastern and western
limits of this deformation. The Great Dyke has been dated at 2580 Ma (Allsopp I965;
Robertson & van Breemen I970) and must have closely followed this ductile deformation. This

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
330 M.P. COWARD

coincidence suggests that the north-northeast trending structures in the greenstones and gneisses
may have influenced the position and orientation of the Great Dyke, or that the Great Dyke
may have been intruded during a period of stress relaxation following this deformation.

DIscUssIo N
Some deformation of the greenstone belts in Rhodesia accompanied the intrusion of the
diapiric granites, but important early phases of deformation pre-dated their intrusion. During
this early deformation there was imbrication and overturning of parts of the stratigraphic pile
and there may have been considerable crustal shortening. However, as intrusive granites
obscure much of this early structure and as there has been intense later deformation, it has not
been possible to place quantitative estimates on the amount of shortening during this early
phase.
This early deformation was followed by regional deformation which produced the main
cleavage in the greenstone belts and granites. Strain measurements indicate a mean shortening
of 49 % across the cleavage in the section between Matsitama and Shabani and the deformation
in the southeastern part of domain 1 was largely by associated simple shear with sinistral sense
of movement.
The northern edges of domains 2 and 3 are also shear zones, of different ages but similar
trend. Together they give the general impression of the east-northeast 'Limpopo trend'. Shear
zones of similar trend occur in central Rhodesia and also in South Africa, where the northern
edge of the zone of granulites is a major fault zone, possibly superimposed on an earlier shear
zone (Mason 1973). Shear zones with an east-northeast trend occur in the Transvaal within the
granulites and also at the northern edge of the Murchison greenstone belt (Graham 1974).
These shear zones were formed by different phases of transcurrent movement, where the
relative motions of the different crustal segments were generally parallel to the strike of the
parally
deformation zones and not across them.

This work was financed by an N.E.R.C. research grant and by funds from the Leeds Institute
of African Geology.

REFERENCES (Coward)
Allsopp, H. L. I965 Rb-Sr and K-Ar age measurements of the Great Dyke of southern Rhodesia. J. geophys.
Res. 70, 977-984.
Amm, F. L. 1940 The geology of the country around Bulawayo. Bull. Geol. Surv. S. Rhodesia35.
Anhaeusser, C. R. 1966 A comparison of pebble and fold deformations in the Nelspruit granite aureole. Econ.
Geol. Res. Unit. Univ. Witwatersrand,30, 1-11.
Anhaeusser, C. R. 1973 The evolution of the early Precambrian crust of southern Africa. Phil. Trans. R. Soc.
Lond. A 273, 359-388.
Anhaeusser, C. R., Mason, R., Viljoen, M.J. & Viljoen, R. P. 1969 A reappraisal of some aspects of Pre-
cambrian shield geology. Geol. Soc. Am. Bull. 80, 2175-2200.
Barr, M. & Coward, M. P. 1974 A method for the measurement of volume change. Geol. Mag. 111, 293-296.
Bickle, M.J., Martin, A. & Nisbet, E. G. 1975 Basaltic and peridotitic komatiites and stromatolites above
a basal unconformity in the Belingwe greenstone belt, Rhodesia. EarthPlanet. Sci. Lett. 27, 155-162.
Cobbold, P. R. 1976 Fold shapes as functions of progressive strain. Phil. Trans. R. Soc. Lond.A 283, 129-138
(this volume).
Cotterill, P. 1969 The chromite deposits of Selukwe, Rhodesia. Econ. Geol. Monogr.4, 154-186.
Coward, M. P., Graham, R. H.. James, P. R. & Wakefield, J. 1973 A structural interpretation of the
northern margin of the Limpopo orogenic belt, southern Africa. Phil. Trans.R. Soc. Lond.A 273, 487-492.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions
ARCHAEAN DEFORMATION PATTERNS IN SOUTHERN AFRICA 331

Coward, M. P. & James, P. R. I974 The deformation patterns of two Archaean greenstone belts in Rhodesia
and Botswana. Precambrian Res. 1, 235-258.
Coward, M. P., James, P. R. & Wright, L. I976 The northern margin of the Limpopo mobile belt, southern
Africa. Bull. Geol. Soc. Am. (In the Press.)
Coward, M. P., Lintern, B. & Wright, L. I976 The pre-cleavage deformation of the sediments and gneisses
of the northern part of the Limpopo belt. In The earlyhistoryof theEarth (ed. B. F. Windley). (In the Press.)
Cox, K. G., Johnson, R. L., Monkman, L.J., Stillman, C.J., Vail, J. R. & Wood, D. N. I965 The geology
of the Nuanetsi igneous province. Phil. Trans. R. Soc. Lond.A 257, 71-218.
Dunnet, D. 1969 A technique of finite strain analysis using elliptical particles. Tectonophysics7, 117-136.
Dunnet, D. & Siddans, A. W. I97I Non-random sedimentary fabrics and their modification by strain.
Tectonophysics 12, 307-325.
Fyfe, W. S. I973 Granites past and present. Geol. Soc. S. Afr. Spec.Publ. 3, 13-16.
Gay, N. C. I969 The analysis of strain in the Barberton Mountain Land, eastern Transvaal, using deformed
pebbles. J. Geol. 77, 377-396.
Graham, R. H. 1974 A structural investigation of the southern part of the Limpopo belt and the adjacent
Kaapvaal craton, South Africa. 18th ann. Rep. res. Inst. afr. Geol. Univ. Leeds.
Hawkesworth, C.J., Moorbath, S., O'Nions, R. K. & Wilson, J. F. I975 Age relationships between green-
stone belts and 'granites' in the Rhodesian Archaean craton. Earth Planet. Sci. Lett. 25, 251-262.
Hickman, M. H. I974a 3500 m.y. old granite in southern Africa. Nature,Lond.251, 296.
Hickman, M. H. 1974 b Geochronological study of the Rhodesian craton and the Limpopo belt, Southern Africa -
a progress report. 18th ann Rep. res. Insi. afr. Geol., Univ. Leeds.

belt at Pikwe, Botswana Bull. ull. Geol. Soc. Am. 86, 1468-1472.
Hossack, J. R. 1968 Pebble deformation and thrusting in the Bygdin area (southern Norway). Tectonophysics
5, 315-339.
Hudleston, P. J. 1973 The analysis and interpretation of minor folds developed in the Moine rocks of Monar,
Scotland. Tectonophysics 17, 89-132.
James, P. R. 1975 A deformation study across the northern margin of the Limpopo belt, Rhodesia. Unpub-
lished Ph.D. thesis. University of Leeds.
Litherland, M. I973 Uniformitarian approach to Archaean 'schist relics'. Nature, Lond. 242, 125-127.
Litherland, M. & Key, R. 1974 A preliminary interpretation of the field mapping of the Eastern Geotraverse
of Botswana. 18th ann. Rep. res. Inst. afr. Geol., Univ. Leeds.
MacGregor, A. M. 1951I Some milestones in the Precambrian of Southern Rhodesia. Trans.Geol.Soc. S. Africa.
54, 27-71.
Mason, R. I973 The Limpopo mobile belt - southern Africa. Phil. Trans. R. Soc. Lond.A 273, 463-485.
Moorbath, S. I976 Geochronology and isotope geology of the early Archaean. In The earlyhistoryof the Earth.
(ed. B. F. Windley). (In the Press.)
Oldham, W. I968 A short note on recent geological mapping in the Shabani area. Tran. Geol. Soc. S. Afr. 71
(annex), 184-194.
Ramsay, J. G. I963 Structural investigations in the Barberton Mountain Land, eastern Transvaal. Trans.
Geol. Soc. S. Africa. 66, 352-398.
Ramsay, J. G. I967 Foldingandfracturingof rocks.New York: McGraw-Hill.
Roberts, B. & Siddans, A. W. B. I97I Fabric studies in the Llwyd Mawr ignimbrite, Caernarvonshire, North
Wales. Tectonophysics 12, 283-306.
Robertson, I. D. M. & Van Breemen, 0. I970 The southern satellite dykes of the Great Dyke of Rhodesia.
Geol.Soc. S. Afr. Spec.Publ. 1, 621-644.
Sohnge, P. G., Le Roex, H. D. & Nel, H. J. 1948 The geology of the country around Messina. Explan.
Sheet 46, Geol.Surv.Dep. Min. S. Afr.
Stowe, C. W. I968 The geology of the country south and west of Selukwe. Bull. Geol.Surv.Rhodesia59, 209 pp.
Stowe, C. W. 1974 Alpine-type structures in the Rhodesian basement complex of Selukwe. J. geol. Soc. Lond.
130, 411-426.
Talbot, C.J. 1970 The minimum strain ellipsoid using deformed quartz veins. Tectonophysics 9, 47-76.
Van Breemen, 0. & Dodson, M. H. I972 Metamorphic chronology of the Limpopo belt, southern Africa.
Bull. geol. Soc. Am., 83, 2005-2018.
Wakefield, J. 1974 The geology of the Pikwe Ni-Cu province, eastern Botswana. Unpublished Ph.D. thesis,
University of Leeds.
Wilson, J. F. 1973 The Rhodesian Archaean craton - an essay in cratonic evolution. Phil. Trans.R. Soc. Lond.
A 273, 389-411.
Wood, D. S. 1973 Pattern and magnitudes of natural strain in rocks. Phil. Trans. R. Soc. Lond. A 274, 373-
382.

This content downloaded from 169.229.32.137 on Thu, 8 May 2014 06:31:05 AM

All use subject to JSTOR Terms and Conditions