Journalof Structural Geology, Vol. 15, No. 1, pp. 1 to 20, 1993 0191-8141/93 $06.00+0.

00
Printed in Great Britain Pergamon Press Ltd

Strain and kinematic analysis in general shear zones

CAROL SIMPSON*

Division of Earth Sciences, National Science Foundation, Washington, DC 20550, U.S.A.

and

DECLAN G . D E PAOR

Department of Earth and Planetary Sciences, The Johns Hopkins University, Baltimore, MD 21218, U.S.A.

(Received 12 December 1991; accepted in revisedform 21 May 1992)

Abstract--We present a unifying theory for the development and distribution of strain markers and kinematic
indicators in zones of general shear, and thus provide a framework in which data previously considered
contradictory may be understood. Rigid and deformable porphyroclast systems, including o, 6 and complex a-~
grains are potential indicators of both strain and flow. The shapes and distribution of such porphyroclast systems
may be used to distinguish among different tectonic regimes.
General shear is divided into two fields: sub-simple shear, in which the rotational component of the strain is less
than that for simple shear, and super-simple shear, in which the rotational component is greater than for simple
shear. Sub-simple shear may involve narrowing or broadening of shear zones. Super-simple shear regimes are
possible in local regions such as the vicinities of deformable porphyroclasts, but must be enclosed by regimes of
sub-simple shear.
The polar Mohr constructions for finite deformation and flow are useful to analyze general shear in theory. The
hyperbolic net is employed for practical plotting of real data and derivation of the kinematic vorticity number,
Wn. This number represents the relative contributions of pure and simple shear in steady flow. In nature,
deformation is thought to build up and decay by processes that may invalidate the assumption of constant flow
regime. We therefore introduce the concept of accelerating deformation and analyze the implications of non-
steady flow for the shearing histories of deformed objects.

INTRODUCTION Mohr diagrams (usually labelled 2, Y or 2', 7'), whereas

infinitesimal extensions e were plotted against engi-
STRUCTURALgeologists possess a variety of tools for the neer's shear strain 7/2.
strain and kinematic analysis of deformed rocks. How- In this work, deformation refers to the transformation
ever, few workers have been able to quantify defor- of material from one geometric or mechanical state to
mation states and paths in the same rocks, partly owing another and shear refers to any deformation that
to the nature of the materials involved. Finite strain is changes the angles between intersecting lines. Finite and
easy to measure in low-grade rocks containing linear or infinitesimal strains are both represented more simply in
angular markers that obey the March model (1932), or terms of stretch S, extension e, and shear strain Y, and
elliptical objects such as reduction spots (Ramsay 1967), the relationship between deformation state and path is
but such rocks rarely contain a record of the strain path. clarified. We show how strain markers, fabric elements
To develop incremental strain markers such as fibrous and objects traditionally used as simple shear sense
beards on pebbles, relatively rigid objects are required, indicators are more easily explained by general shear in
but such objects do not accurately record finite strain. a high strain zone. In addition, these markers yield
Rotated structures such as dynamically recrystallized information about the velocity gradients leading to that
porphyroclasts are common in high strain mylonites with state of general shear, and the accelerations involved in
a strong fabric (Simpson & Schmid 1983, Hanmer & the waxing and waning stages of the flow.
Passchier 1991), but these rocks often lack cumulative or
even incremental strain markers. As pointed out by
Passchier (1986) and re-emphasized by Choukroune et GENERAL SHEAR
al. (1987), the quantitative relations between non-
coaxiality and structural asymmetry have remained un- Deformation state
certain. Furthermore, theoretical relationships between
deformation state and path have been obscured in the Since the early work of Sorby (1855), Harker (1885)
past by measurement in different reference frames. and Wettstein (1886) structural geologists have tended
Finite strains were traditionally plotted on quadratic to rely on three simplistic deformation models: pure
shear, simple shear and rigid rotation. These terms have
*Also at: Department of Earth and Planetary Sciences, The Johns strict definitions in classical mechanics but are used more
Hopkins University, Baltimore, MD 21218, U.S.A. casually in structural geology. Pure shear is used to
2 C. SIMPSONand D. G. DE PAOR

denote any deformation whose three principal stretch- polar co-ordinate reference axis or zero direction (Fig.
ing directions are eigenvectors (directions of no ro- lb). In this reference frame, the shear zone boundary is
tation). Simple shear is a deformation whose clearly a direction of no rotation, i.e. an eigenvector. To
eigenvectors are confined to a plane--the shear plane. construct the Mohr circle, the baseline stretch vector ~1
For rigid rotation, there is just one eigenvector--the in geographical space is rotated through 90° and marked
rotation axis. along the reference axis. This point has polar co-
Ramsay & Graham (1970) demonstrated that simple ordinates (~l, 0) in strain space. The deformed normal to
shear is the only constant volume strain regime that can the shear zone is marked at (So, ~p); being the only point
occur in straight, parallel-sided shear zones bounded by representing a stretched line simultaneously in geo-
undeformed wall rocks unless there is a discontinuity graphical space and strain space, it is termed the anchor
(fault) along the boundary. More general deformation point. The circle is drawn diametrically through (~1, 0)
states were first discussed by Ramsay & Wood (1973) and (So, ~P) (Fig. lb). In the special case of irrotationai
and Sanderson (1976), who examined the superposition strain this circle's center falls on the reference axis.
of plane strain on an initial uniaxial compaction. Mat- The stretch and rotation of a particular material line
thews et al. (1974) factorized deformation into pure are represented by a point (S, a) on the perimeter of the
shear and orthogonal simple shear along bedding circle (Fig. lc). Given this point, the corresponding line
planes. Other work on general deformation includes: orientation in geographical space is obtained by pivoting
curved shear zones (Coward 1976); dikes that intruded the stretch vector about its second intersection with the
during combined shear and extension (Escher et al. circle until it passes through the anchor point (Fig. lc).
1975); non-coaxiality in progressive deformation This gives the line's final orientation fl' in geographical
(Means et al. 1980); and shear zones with non-parallel space. Its initial orientation fl differs from fl' by a, the
sides and terminations (Simpson 1983). rotation of the line.
De Paor (1983) defined two fields of general shear-- The points on the circle that are furthest from and
namely sub-simple shear, in which the rotational com- nearest to the origin represent the lines of maximum and
ponent of strain is less than that for the simple shear of minimum stretch, $1 and S2. They were initially 90 ° apart
the same intensity, and super-simple shear, in which the in geographical space and, having undergone equal
rotation is greater than for simple shear. Some authors rotations ~o, they are also orthogonal after deformation;
ignore the latter category and use the term general shear hence they have suffered no net shear strain (Fig. ld).
for the former. Although super-simple shears are gener- Any other pair of diametrically opposite points on the
ally confined to special places such as the interiors of circle represents the stretch and rotation of a pair of
deformable porphyroclasts or regions of particulate initially orthogonal lines in geographical space and the
flow, they may constitute an extremely important com- angle they subtend at the origin is their angular shear ~p
ponent of heterogeneous deformation regimes (Talbot (Fig. le). Maximum angular shear is recorded by the
& Jackson 1987). pair of lines that were initially oriented symmetrically at
Sub-simple shear zones may become narrower or ___45° to the principal directions (Fig. lf). Their final
broader. Reduction of a shear zone's thickness may be orientations are indicated in Fig. l(g). In Fig. l(h) the
achieved by (a) increase in length, (b) reduction in two points where the circle intersects the reference axis
volume, or (c) extension in the third dimension (i.e. represent the eigenvectors (~1, 0) and (~2, 0) of this
transpression, Sanderson & Marchini 1984). In this general shear. The stretches ~l and ~2 are the corre-
paper we mainly consider the temporal evolution of a sponding eigenvalues. The corresponding directions in
parallel-sided shear zone's profile plane, but much of the geographic space are obtained by the pivoting manouver
discussion may be applied to spatial gradients in shear described above which yields the shear zone boundary
zones that have non-parallel sides. and a second direction inclined in the displacement
direction (Fig. lh). Note that this figure illustrates the
Mohr circle for strain case of a narrowing sub-simple shear zone. In the
opposite case of a broadening zone, the eigenvalue ~l
The Mohr circle construction for strain serves as a would be less than ~2-
powerful tool for studying general shear. The traditional Different sign conventions for Mohr constructions are
Mohr construction (Brace 1961) employed a quadratic advocated by various authors (see summary in Allison
2', 7' reference frame and was limited to irrotational 1984). There are eight possible choices of sign conven-
strains. The construction for strain introduced by De tion; we prefer the one described above because: (1) it
Paor (1983) is simple and very similar in principle to the reduces the construction steps to a minimum; (2) the
well-known Mohr circles for stress. In the stress case, the anchor point links strain space to geographic space; (3)
Cartesian co-ordinates of each point on the circle rep- rotational strains are measured in the same sense in
resent the normal and shear stresses acting on a certain strain space and geographical space; and (4) stretches
plane, whereas in De Paor's construction for strain the and rotations of lines are referred to the final state
polar co-ordinates of a point on the circle represent the preserved in the rock. By comparison, the polar co-
stretch and rotation of a certain line (Fig. la). For ordinates of Means' (1982, 1983) construction are the
general strain in shear zones, the zone boundary is the stretch and the negative of the rotation of a line. His
baseline and the normal to the shear zone is then the circle's pole, which plots diametrically opposite our
 reference
(a) (b) axis $1 reference
(C) a ~ [3 /T
~ (~,o)
g (~2, :so ,~
stre
n°rrna'°° /
o ' ,, a e,ne

~.. " , ' / / / ~ H / / H / / / H / ~ / ,

.~,,,-,,,,,,~,,~,~g l ~ / / / / / / / / / / / / / / / / . ~
~o

(d) s~ (e) (f) (g) (h)

(~, 0)
i f ,.7

0
(~2,
S2
% /
y=O
N
o
Fig. 1. Explanation of construction for strain. (a) Traditional Cartesian Mohr construction for stress, amd similar polar construction for strain. Vertical line is
polar reference axis. (b) Baseline is shear zone boundary (shaded). Its normal is polar reference axis. Parallelogram represents deformed original square. ~l is
stretch of shear zone boundary; ~ ~,0 represents baseline in Mohr space. So and ~p are stretch and rotation of pole to shear zone. Solid square is anchor point: it
represents deformed pole to shear zone in real space and in Molar space. ~2 is second eigenvalue. S I and $2 are principal stretches. (c) Given point (S,a) on
he Mohr circle, corresponding direction in real space is obtained by taking line through S and pivoting about its second intersect point on the Mohr circle (open
circle) until it passes through anchor point (solid square). This is line's deformed orientation/3'. Initial orientation is obtained by dashed construction lines. Sets
of lines represent undeformed (dashed) and deformed (solid) directions. (d) Principal stretches S I and S 2 plot on circle's central axis. o~ = rigid rotation;
angular shear is zero. (e) Angular shear qJ is given by angle subtended at origin by any diametrically opposite pair of points. (f) Maximum W occurs when
triangle is isosceles. (g) Construction for maximum shear ~max (= tan ~Pm~×). Geographical directions as in (f). Initial directions, i, are perpendicular; final
directions, f, are not. A = anchor point. Construction follows rules outlined in (c). (h) Lines of zero rotation in geographical space and Mohr space, v is angle
between eigenvectors. ~ , ~2 are corresponding eigenvalues.
4 C. SIMPSONand D. G. DE PAOR

anchor point, identifies the initial orientation of the other begins perpendicular to it and progesssively ro-
corresponding direction in geographical space. tates toward the flow direction. The long axis of the
strain ellipsoid always bisects the acute angle between
Deformation path the circular sections, so it too rotates toward the flow
direction.
Deformation is a progressive phenomenon that fol- The rate of rotation of lines increases with obliquity to
lows a particular deformation path. The path is a closed the flow plane and reaches a maximum in the perpen-
loop in stress-strain space in the case of elastic behavior; dicular plane. Once lines pass through this plane, their
alternatively an open-ended deformation path may lead rate of rotation slows down so they approach but never
either to fracture and rupture of the material, or to reach the flow plane. The rotation rates of the strain
permanent strain. Ramberg (1975) derived particle ellipsoid axes also decelerate, although not as fast. In
paths for various combinations of coaxial and non- simple shear, material lines which lie along the S~ and $3
coaxial flow. Particle paths represent the cumulative axes of the strain ellipsoid rotate faster than the strain
changes in position of points during progressive defor- ellipsoid axes; thus material constantly flows through
mation that is built up by identical increments of infinite- two principal planes of the strain ellipsoid. After passing
simal strain (Fig. 2). A consequence of this time through the principal plane containing St, a material
invariance is that the displacement of a particle is a line's rotation rate is reduced and soon becomes slower
function of position only. When one particle reaches a than the rotation of the principal axes. Thus the angle
position previously occupied by another, its displace- subtended between a material line and the immaterial
ment is identical to that of its predecessor. Ramberg long axis of the strain ellipsoid passes through a maxi-
(1975) showed that particle paths could be open-ended mum value and then reduces. The implications of this
or form closed loops. Weijermars (1991) illustrated theoretical result for practical analysis of fabrics are
Ramberg's particle paths for the special case where significant. A small rigid elongate crystal will tend to
stress and infinitesimal strain axes are coincident (i.e. grow fibrous tails as it rotates in progressive simple
where there is no material anisotropy). shear. These tails are deflected towards the S~ direction
and so their shape varies with time in a complex way. In
Progressive stretch histories and rotations of lines super-simple shear, all lines rotate in the same sense but
at different, pulsating rates.
The stretching histories of lines during progressive Planes' are rotated and deformed in a complex way
shear are complex (Flinn 1962). Ramsay (1967) showed during progressive deformation (March 1932, Owens
that lines oriented close to the S~ direction may lengthen 1973). Even in progressive pure shear, deformation
throughout their deformational history whereas lines involves rotation in any plane that is oblique to all three
close to the $3 axis may shorten first and then lengthen. principal axes. Distortion is accompanied by rotation of
The rotation histories of lines during general shear are the plane's pole and by spin about that pole. Even if
also complex. In progressive pure shear, material lines volume is conserved, areas of oblique planes, which may
tend to rotate away from $3 towards S~, veering toward be calculated using De Paor's (1983) tensor, are pro-
$2 en route (Flinn 1962). Rotation increments are gressively altered. Some are shrunk, then restored, and
greatest at a high angle to all three principal directions finally spread beyond their initial area. Poles to planes of
and drop to zero along each principal direction. In sub- no area change form loci similar to directions of no
simple shear, there are two unequal fields of opposite longitudinal strain.
rotation, bounded by the eigenvectors. Lines parallel to During shear deformation, lines oblique to the princi-
the eigenvectors do not rotate. Simple shear causes all pal directions rotate out of orthogonality to their cross-
material lines oblique to the flow plane to rotate toward sectional planes and thus become sheared. After a finite
the flow direction. One circular section of the simple simple shear, some directions are sheared dextrally and
shear strain ellipsoid is attached to the flow plane, the others sinstrally (Fig. 3). Furthermore, during defor-
mation some lines' shear strains progressively increase
in magnitude whereas others decrease. Eventually, each
(b) sinistrally sheared line in a dextral shear zone attains a
maximum shear value, reduces its shear strain to zero as

increasing max . reasing

T ~clec
• zero..~ zero
decreas'~/~ ~ ~ - ' ~ increasing
max ,.,..- -'---,,Jr "."max
Fig. 3. History of shear strain of individual lines during progressive
Fig. 2. Particle paths, after Ramberg (1975). (a) Pure shear, (b) dextral simple shear. 'T's represent initiallyorthogonal lines. Depend-
sub-simple shear, (c) simple shear, (d) super-simple shear, (e) rigid ing on orientation, shear strain may be increasing or decreasing in
rotation. current increment of deformation.
 Strain and kinematic analysis in general shear zones 5

it passes through the long axis of the strain ellipsoid and el and g2, plot on the infinitesimal Mohr circle's vertical
then becomes sheared dextrally. Similarly, some dex- diameter (see Fig. 4c). They share the same angle of
trally sheared lines reduce their shear strain to zero as rotation ~, so they are therefore directions of no infinite-
they pass through the short axis of the strain ellipsoid simal shear strain. All other pairs of orthogonal direc-
and then become sheared sinistrally. Thus it should not tions plot diametrically opposite on the Mohr circle and
be surprising that individual kinematic indicators in a the difference between their a co-ordinates represents
materially heterogeneous rock show shear sense their angular shear ~ and, because small angles approxi-
opposite to the bulk shear direction. In progressive pure mate to their tangents, the angular difference equals the
shear, the shear strain builds up continuously and mon- infinitesimal shear strain, 7. Horizontally opposite
otonously; unlike the case of simple shear, there are no points on the circle represent directions of maximum
turning points in the magnitude of shear strain. For sub- shear strain, which are also directions of mean extension
simple shear, there is one turning point. Shear strains e. When the Mohr circle is centered on the origin, there
pulsate for super-simple shear. is no infinitesimal rotation or dilatation; co-ordinates of
any point on its perimeter are then the components of
Mohr construction for infinitesimal strain longitudinal strain and 'engineer's shear strain', e and
7/2 (Fig. 4d).
In the case of infinitesimal strain, the Mohr circle is The points U1 and U2 farthest from and nearest to the
centered near the unit stretch (S = 1.0) position on the origin on the Mohr circle for infinitesimal strain rep-
reference axis, illustrated schematically in Fig. 4(a). A resent the axes of an ellipse termed the displacement
many-fold enlargement of this circle is shown in ellipse (Fig. 5) (De Paor 1983). The displacement ellipse
Fig. 4(b). Stretch, S, and extension, e, are related as is comparable to the strain ellipse in that it represents a
S = l + e , so the vertical axis represents extension general tensor quantity (the displacement gradients ten-
measured as +e and - e relative to a new origin at sor), but it is also comparable to the stress ellipse in that
S = 1.0. Rotation angles are still measured about the its principal axes need not be of the same sign. One axis
S = 0 origin. Angles measured in radians around a curve may represent an outward directed displacement and
are defined in terms of arc-length per unit radius, and the other an inward directed displacement, as in the case
very small arcs approach straight lines, thus radian where the Mohr circle encloses the origin.
values plot approximately along the horizontal axis The displacement ellipse is an important concept in
through the new origin at e = 0. The old polar reference progressive deformation studies. For non-coaxial defor-
frame can be approximated, therefore, by a Cartesian mations, the directions of maximum and minimum dis-
system with e as ordinate and a as abscissa (Fig. 4b). placement are not parallel to the directions of maximum
The directions of maximum and minimum extension and minimum extension. For example, maximum dis-
placement parallels the boundary of a simple shear zone,
but maximum extension is at 45 ° to the boundary.
(S,ct) Furthermore, the directions of maximum displacement
(a)
and maximum incremental shear strain are both oblique
to the boundary of a general shear zone.
........... iraclians) Velocity gradients and flow
t The term flow is widely used in continuum mechanics
to denote the process of permanent, non-elastic, or
time-dependent strain. Flow is thus the instantaneous
(c) E rate of infinitesimal strain. Flow regimes are coaxial if
E1 successive increments of infinitesimal strain are parallel
U1~ O. and steady if they are also of equal magnitude. Time
derivatives of e and a in the Mohr circle construction for
~/max1 ~max...... infinitesimal strain plot as strain rates, k, along the

E2 initialcircle
Fig. 4. Relationship between finite and infinitesimal Mohr construc-
tions. (a) For very small strains, finite Mohr circle (in S-a reference
~ d i s ~ p displacementellipse
frame) is infinitesimal (exaggerated here for clarity). Arrow depicts
magnification of view of top left corner of slightly deformed unit
square. (b) Dotted line is a new a reference line at distance of one unit
up from old (S=0) origin. In new reference frame, co-ordinates are
extension e and rotation in radians, a. (c) Positions of el, e2, 7max (see
lace's
text). U 1, U 2 represents axes of the displacement ellipse (see Fig. 5).
(d) Mohr circle centered on origin. There is no infinitesimal rotation Fig. 5. The displacement ellipse on right is the locus of all displacement
or dilatation; co-ordinates of any point on its perimeter are e and vectors joining points on the initial unit circle to the corresponding
y/2 (= a/2). The latter is justified because small angles measured in points on the strain ellipse on left. Note that some displacement
radians approximate their tangents. vectors point outwards and others inwards.
/
6 C. SIMPSONand D. G. DE PAOR

~2
Fig. 6. The flow Mohr circle, kl, kz are directions of maximum and
minimum extension rate and share the same angular velocity gt. UI
and 1-)2are displacement rate axes.

vertical axis and angular volocities, t~, along the horizon- Fig. 7. Schematic representation of gradient, divergence and curl.
tal axis (Fig 6). Division by time modifies the reference
frame without affecting geometrical relations in Mohr
space; the Mohr circle now represents flow velocity while it is deforming. Vorticity is recorded by thehori-
gradients at an instant in the deformation history. zontal offset of the center of the flow Mohr circle, gt. It is
The flow ellipse is the time derivative of the displace- an unfortunate fact of history that vorticity in classical
ment ellipse. It is important conceptually because it mechanics is defined in terms of a mathematical function
highlights the fact that there is no unique flow plane in called the curl of the velocity fi__eld, and is twice_the
general shear. In progressive simple shear, the flow physically meaningful quantity, 4. We here define & as
Mohr circle passes through the origin (k = 0). The flow the geologist's vorticity. Engineer's vorticity, defined as
ellipse's short axis is of zero length, and its shape the velocity field's curl, is thus twice geologist's vorticity,
degenerates to a line representing the flow direction. 24. This is inversely analogous to the case of geologist's
The flow Mohr circle shows that simple shearing is not shear strain y vs engineer's shear strain y/2. Vortical flow
just a vectorial process, but rather is a tensor phenom- is applicable, for example, to deforming and rotating
enon. clasts within a matrix of less vortical flow regime.
The directions of maximum and minimum extension
rate, kl and k2, plot on the flow Mohr circle's vertical
diameter (Fig. 6). They share the same angular velocity MEASUREMENT OF VORTICITY
4, so they maintain orthogonality in the current incre-
ment of deformation and are therefore directions of no To evaluate vorticity in rocks, one may choose from
shear strain rate. Although material in these directions three potential reference frames: the finite strain axes
does not shear, it does rotate at the mean rate of ~. For (Elliott 1972); the infinitesimal strain axes (Lister &
steady flow conditions, the kl and k2 directions are fixed Williams 1983); or the shear zone boundary (this work).
in space as material sweeps through them. Infinitesimal axes are usually chosen in theoretical work
(e.g. Lister & Williams 1983), based on the argument
Stretching, shearing, divergence and curl that fabrics are sensitive to the instantaneous directions
of maximum extension. However, in practice the geolo-
A ductile flow regime possesses three important attri- gist often has to measure vorticity relative to the only
butes: gradients, divergence and curl (Schey 1973), available visible markers, namely, the shear zone
represented schematically in Fig. 7. Velocity gradients boundary and its normal for an external reference
(grad) lead to stretching of the distances between par- frame, and the foliation plus the normal to foliation for
ticles, rotation of lines that join them, and shearing of an internal reference frame. Provided the reference
poles to planes. Divergence (div) denotes the tendency frame itself is not rotating--a phenomenon called spin
for isotropic expansion or contraction of material about (Truesdell 1954, Lister & Williams 1983), it does not
a point in the flowing mass (see Talbot & Jackson 1987). matter which is chosen; Fig. 6 shows that the mean rate
The vertical offset of the center of the flow Mohr circle, of rotation of all pairs of perpendicular lines is the same.
the mean longitudinal strain rate ~-, represents the However, if the shear zone's wall rock is deforming,
divergence, which may be centrifugal or centripetal. The then only that portion of the vorticity which is internal to
Laplacian is a measure of the gradient in the magnitude the system is measured. The internal vorticity of a
of the divergence in a flowing mass. Divergent flow and flowing mass is strictly defined as the rate at which
Laplacians are relevant to the understanding of high material sweeps through the infinitesimal extension axes
strain zones with non-parallel boundaries such as el, e2. We may not know whether a particular fabric
spreading thrust nappes. tracks the finite or infinitesimal principal directions so
Vorticity is a description of the way material rotates internal vorticity may be difficult to quantify. However,
 Strain and kinematic analysis in general shear zones 7

the sense of rotation is all that is measured in practice al. 1987), but quartz fabrics appear to be more com-
and so the choice of foliation as an ad hoc internal plexly related to Wn (Schmid & Casey 1986).
reference frame should not present a problem.
Acceleration
Kinematic vorticity number
In rigid-body mechanics, it is not sufficient to describe
Internal vorticity is positive or forward directed when the position and velocity of a particle; for a full descrip-
material sweeps through the reference axes in the direc- tion its acceleration is also required. Similarly, finite
tion in which they are themselves rotating, otherwise it is strain and flow are only a partial description of a rock's
negative or backward. For positive cases, it is useful to deformation; acceleration completes the story. In kine-
further classify vortical flow regimes in two dimensions matic studies it is usual to assume, for the sake of
by the amount of vorticity relative to the amount of simplicity, that the flow regime remains constant with
longitudinal strain. The commonest measure of this time, but this is unlikely to be true in general. Changes
ratio is the kinematic vorticity number Wn (Truesdeil are inevitable at the start and end of a tectonic event and
1954), which can be defined in several ways; most where the boundaries of a high strain zone are curved.
simply, it is the cosine of the angle v between the Acceleration is critically important in kinematic studies
eigenvectors (Bobyarchick 1986) (Fig. 8), because the consequent change in strain rates may alter
the operative deformation mechanisms. Microstruc-
W. = cos (v). (1)
tures attributed to a change in P-Tenvironment may be
In the case of progressive simple shear, Wn = due instead to acceleration or deceleration of defor-
cos 0 = 1.0. This is consistent with the fact that the flow mation. Furthermore, we do not know whether it is the
Mohr circle touches the reference axis at one point only. maximum displacement rate (long axis of the flow
In the case of sub-simple shear, W. has a value between ellipse), maximum extension rate (vertical coordinate
0.0 and 1.0; correspondingly there are two non- on the flow Mohr circle), maximum shear strain rate
orthogonal eigenvectors of the flow Mohr circle. (diameter of the flow Mohr circle) or other quantity
According to Means et al. (1980), petrofabric devel- (e.g. mean displacement rate) that controls the choice of
opment may be controlled by the kinematic vorticity deformation mechanisms.
number of the flow immediately prior to the fossilization If velocity gradients change with time, the flow be-
of the fabric. This relationship has been demonstrated to comes unsteady or accelerating and the flow Mohr
hold for calcite lattice preferred orientations (Wenk et circle's size and position is a function of time. An
accelerating deformation is represented by an acceler-
ation Mohr circle in ~/6~space (Fig. 9a). Change in the
divergence of the flow is recorded by the vertical offset
of the acceleration Mohr circle along the ~ axis. This
means that the rate of dilatation, or rate of inflation, is
speeding up or slowing down. A horizontal shift of the
Mohr circle's center represents a spinning deformation
in which the principal stretching directions change with
time. The rate of spinning is given by the difference
between internal vorticity and total vorticity &, per unit
time.

~///////~ Acceleration and rotation reversals

I ~~ (b)
For steady sub-simple flow regimes, the eigenvectors
of finite strain and flow are parallel. These eigenvectors
are not merely directions of no net rotation, rather they
are directions which have never rotated throughout
deformation. This distinction is important for fabric
considerations. In nature, it is likely that either one or
both of the components of a sub-simple shear will grow
or decay with time. For example, a constant progressive
simple shear in an overthrust orientation may be accom-
panied by a gradual build-up of vertical pure shear due
to the advance of an overlying thrust sheet, or by the
decay of pure shear due to erosional removal of over-
burden. Alternatively, overburden may generate a
Fig. 8. Mohr and geographical space representations of kinematic growing coaxial spreading flow in a thrust sheet during
vorticity number W n = cos v, where v is angle between eigenvectors
(after Bobyarchick 1986). Dextral sub-simple shear zone in (a) is
the commencement of ramping and accompanying shear
narrowing and in (b) is broadening. deformation. The waning stages of a deformation event
8 C. S|MPSON a n d D . G . D E PAOR

(a) (b) (c) ~ (d)

Fig. 9. Representation of acceleration using Mohr circles. (a) Vertical offset of circle on i//i axes represents an inflating
deformation in which there is a change in divergence. Horizontal offset represents a spinning deformation in which there is a
change in the principal flow directions. (b)-(d) Flow Mohr circles for the same cumulative deformation but with different
flow histories. Only the reference diameters of early Mohr circles are drawn for the sake of clarity. (b) Constant flow regime
in which eigenvectors, and thus reference diameters, remain parallel. (c) Coaxial component of flow builds up faster than
the vortical component. (d) Vortical component builds up faster than non-coaxial flow component.

may also involve different rates of decline in the pure simple shear, rigid rotation and dilatation during each
and simple components of progressive general shear. increment of time must be evaluated. This information is
The consequences of accelerating flow regimes are illus- essential for meaningful pressure-temperature-time-
trated using the series of Mohr circles in Figs. 9(b)-(d). deformation paths. Finite strain markers record the
There are three degrees of freedom which can be rep- geometry of the final state only. Kinematic analysis is the
resented by the changes in center and radius of the search for indicators of the flow regime operative during
circle. For the sake of clarity, only one 'reference' deformation. Of particular interest are fabrics that re-
diameter of smaller Mohr circles is shown. The full circle veal the sense of vorticity of the flow. In special circum-
is shown for the last state recorded. stances, it may be possible to evaluate the kinematic
Figure 9(b) illustrates the case of a constant flow vorticity number.
regime. Because the eigenvectors must remain constant One of the oldest concepts of structural geology is the
in orientation, successive reference diameters of the symmetry principle of Curie (1894). It implies that
Mohr circles are parallel. In Fig. 9(c) the build-up of the symmetric fabric patterns result from coaxial progress-
coaxial component of flow exceeds that of the non- ive deformation of a random protolith fabric, whereas
coaxial component. Successive reference diameters are asymmetry may be expected to result from a vortical
more steeply inclined, implying a larger angle between deformation path. The association between fabric sym-
the eigenvectors. In contrast, Fig. 9(d) shows the grad- metry and deformation path is not, however, that simple
ual decrease in this angle resulting from a greater build- (Choukroune et al. 1987). It is important to distinguish
up of non-coaxial flow. The final state is identical in each between passive and active strain markers, between
case. Note that all directions oriented between the first finite and infinitesimal markers, and between material
increment and final cumulative reference diameters and non-material lines.
have undergone a transient reversal in the signs of their
angular velocities. The result is recovery of initial orien- Foliation as a strain m a r k e r
tations followed by further rotation. Lines oriented
between the final cumulative and final incremental refer- The simplest method for the determination of strain in
ence diameter have reversed their sense of rotation, but a shear zone assumes that (i) the cross-sectional trace of
have not yet recovered their initial orientations (Fig.
10). This situation is reminiscent of the shear of lines as a
function of orientation (Fig. 3), and the lengthening of
lines that initially shorten.

" foliation
PRACTICAL ANALYSIS OF GENERAL SHEAR
(a)
In order to determine the complete deformation his- ""..
tory of a rock, the relative contributions of pure shear, baseline

(b) /. . . - ~

Fig. 11. Relationship between foliation direction tp' and shear strain y
Fig. 10. Reversal of rotation sense of lines in cumulative deformation. in (a) simple shear zone, (b) zone of sub-simple shear. Dashed lines are
Arrows indicate the direction of rotation. See text for discussion. traces of circular sections of the strain ellipsoid.
 Strain and kinematic analysis in general shear zones 9

foliation represents the long axis of the strain ellipse q~'

and (ii) the strain is ideal simple shear (Fig. 11a). The
(a) (b)
equations for simple shear are:
r.~x~~ r.~
7 = 2/tan 2~' (Ramsay & Graham 1970) (2)
S1 = cot ~' (3)
$2 = tan ~O' (4)
and Fig. 12. (a) Relationship between the eigenvectors of sub-simple shear
in Mohr space and geographical space. Equation (7) follows from this
7 = 2 sinh E, (5) geometry. (b) Proposed relation between synthetic (C') and antithetic
(C~) shear band orientations and directions of maximum shear strain
where rate ~.
E = (1/2) In ($1/$2) (Nadai 1963). (6)
Because the angle q~' between the foliation and shear
zone boundary is measured in the rock, these calcu-
plane of the simple shear component, but the other
lations can be carried out rapidly on field location. The
eigenvector has an orientation that is dependent on the
main drawback is the assumption of ideal simple shear.
pure shear and simple shear contributions to sub-simple
If the deformation is a sub-simple shear as indicated in
shear. The smaller the irrotational component, the
Fig. l l(b), the method is invalid and angle q~' may
smaller is the angle between ~1 and ~2. The important
become very small at quite low strain intensities.
conclusion is that determination of the eigenvectors in a
general shear zone may lead to information about the
Asymmetry of SNFLS and SNILS relative contributions of irrotationai vs rotational strain.
Much recent effort has centered on ways of identifying
One way to determine the relative contributions of
the eigenvectors (Ramberg 1975, Ghosh & Ramberg
pure and simple shear during steady flow is to find the
1976, Passchier 1987).
surfaces of no finite longitudinal strain (SNFLS) and
compare these with known surfaces of no incremental
longitudinal strain (SNILS; de Sitter 1956, Flinn 1962, Shear bands
Ramsay 1967, Talbot 1970, Passchier 1988). Ramsay
(1967) showed that in progressive pure shear, the Shear bands have been used by a number of workers
SNFLS and SNILS are coaxial. Lines that progressively to obtain sense of shear (e.g. Weijermars & Rondeel
shorten (folds), lines that originally shorten and then 1984, Bobyarchick 1986, Dennis & Secor 1987).
lengthen (boundinaged folds), and lines that progress- Bobyarchick (1986) suggested that the inclined eigen-
ively lengthen (boudinage), are developed symmetri- vector ~2 may represent the orientation of shear bands in
cally about the principal axes. For progressive simple natural shear zones. If correct, then the measured angle
shear and sub-simple shear, SNFLS and SNILS are between shear bands and the overall flow plane could be
oblique to each other so that there are two asymmetri- used to estimate the relative contributions of pure and
cally developed zones of boudinaged folds; the orien- simple shear in natural examples. However, the eigen-
tation distribution of these can be measured in the field vector directions are not directions of maximum shear
and can distinguish between a predominantly pure- strain (in fact, they may undergo quite low angular
shearing or a simple-shearing history. The relative size shears) and the inclined eigenvector is an unstable
of the asymmetrical zones may even permit an estimate direction where there is extension parallel to the shear
of the relative contributions of the two. For progressive zone boundary and shortening perpendicular to it. If a
super-simple shear, there is a zone of lines that lengthen plane is deflected slightly from this direction, it will
first and then shorten, resulting for example in the continue to rotate away progressively.
folding of previously boudinaged veins. Conjugate shear bands (extensional crenulation
cleavage) in zones of non-coaxial strain have been docu-
Orientation of eigenvectors mented by several workers. Platt & Vissers (1980)
suggested that conjugate sets might indicate a coaxial
In Fig. 12(a), polar co-ordinates (~1, 0) and (~2, 0) on flow regime, but subsequently Platt (1984) showed that
the Mohr circle represent the two non-rotating lines conjugate sets could occur if there was deviation from
(eigenvectors) in the rock and the eigenvalues ~t and ~2 progressive simple shear. Harris & Cobbold (1985)
are the principal values of the irrotational strain com- showed the development of conjugate shear bands dur-
ponent. T is the transverse component of deformation. ing bulk simple shearing if sliding was permitted along
The angle 2v, measured around the arc between ~1 and pre-existing foliation planes. Behrmann (1987) reported
~2 in Mohr space, provides the angle v between these the occurrence of conjugate sets of shear bands in a
two eigenvectors in real space, where mylonite from the base of a known thrust nappe. The
mylonite gave a consistent shear sense using quartz
tan (v) = (~i - ~2)/T. (7) petrofabric data, and was thought to be developed in a
The orientation of one eigenvector is fixed to the flow non-coaxial (sub-simple shear) regime. One shear band
10 C. SIMPSONand D. G. DE PAOR

set was inclined at 30° towards the flow direction and

sheared synthetically; the other was inclined at 60° I=-
infinitesimal strain fabric finite strain fabric w
upstream and sheared antithetically. These angles
suggest a possible association with the directions of
maximum shear strain rate (Fig. 12b) and not with the / /
//
eigenvector directions.
//
In view of the alternative interpretations, shear band
/
orientations are, at best, semi-quantitative indicators of / ,/ / / /
the shear components. If shear bands approximately
bisect the angle between the eigenvectors in zones of
relatively low strain, then they will become progress- (a) Simple shear (b) Sub-simple shear
ively rotated into alignment with the flow plane at higher Fig. 14. Illustration of possible ambiguities arising from spiral in-
strains. clusion trails in rigid porphyroclasts such as garnets. (a) Conventional
trails result from garnet rotating more rapidly than foliation. (b)
Relatively rapid rotation of foliation during sub-simple shear may be
Rotation of rigid spherical objects wrongly interpreted,

Rigid quasi-spherical objects such as garnet crystals or

pyrite framboids are excellent kinematic indicators rigid porphyroblast in simple shear (Fig. 14a). This is
(Rosenfeld 1968, Schoneveld 1977), but the relationship self-evident for a steady state foliation, which by defi-
between internal and external fabrics may be complex. nition does not rotate. Where the foliation tracks the
From the traditional interpretation of snowball garnets finite strain axes, it's rotation rate decelerates as it
(e.g. Schoneveld 1977), the porphyroblast overgrows approaches parallelism with the shear zone boundary
the existing foliation as the garnet rotates. The internal and so the total rotation of the rigid grain must exceed
fabric is dragged around as the garnet rotates faster than that of the foliation. If the shear zone undergoes sub-
the foliation. In progressive simple shear the rigid simple shear and broadens with time, then the foliation
rotation rate for rigid objects is half the instantaneous remains at a high angle to the zone boundary and
shear strain rate (Fig. 13): enhances the rotational pattern of inclusions. However,
if there is a component of pure shear acting to narrow the
- 7 (8) shear zone and the foliation tracks the finite strain axes,
2 then the foliation may rotate faster than the rigid object
This is because half of the angular shear experienced by provided the latter does not rotate more than 90°. In a
the pole to the shear zone is a result of rigid rotation and domain of super-simple shear that includes a rigid
the other half is caused by the pure shear at 45° to the object, a foliation which tracks the finite strain direction
shear zone boundary. In progressive pure shear, the may rotate faster than the object at all times, resulting in
matrix flows symmetrically around rigid obstructions a false interpretation of the sense of shear even at high
and applies equal fiber loading stress to each side, so the strain. Interpretation of snowball garnets thus relies
object does not rotate. Applying the same arguments to heavily on the correct interpretation of the relationship
general shear, rigid spheres such as garnets are unaffec- between internal and external fabrics.
ted by the pure shear component of deformation and Curvature of fibers around rigid inclusions can be used
rotate only in response to the rigid rotation component to obtain the rotational component of strain provided
(Ghosh & Ramberg 1976). there is a clear distinction between face-controlled fibers
At low finite strains, foliation lags behind a spherical and those that track the incremental stretch direction
(Ramsay 1980, Ramsay & Huber 1987). A difficulty
arises, however, if the fibers do not form parallel to the
incremental stretch direction (Urai et al. 1991).

Rotation of rigid elongate objects

Reverse rotations of objects are generally possible in

homogeneous flow regimes only if the deformation is
sub-simple shear, the objects are relatively rigid, and
they are more elongate than a critical value. As shown
CO by Gay (1968) and Ghosh & Ramberg (1976), there is a
Fig. 13. Simple shear may be factored into a component of rigid simple relationship between the rotation angle of rigid
rotation through angle to whose tangent is half the shear strain, and a spherical objects and that of elongate objects in general
pure shear component oblique to reference axes. The latter com-
ponent adds to the rotation of shear zone's pole but cancels the zone shear zones with constant flow parameters, i.e. where
boundary's rotation. A circular rigid object responds to rigid rotation the eigenvectors remain constant throughout. The more
to but not to the irrotational component of deformation. Note that this elongate objects are more susceptible to the pure shear
argument may be applied to every increment of deformation and thus
holds even where actual deformation path is not that represented by
component and tend to rotate both forwards and back-
arrows. wards. By measuring the angle of rotation of rigid
 Strain and kinematic analysis in general shear zones

Fig. 15. (a) Complex o-6 feldspar grain from mylonitic granodiorite, Pinalefio Mountains, Arizona. Both o and b tails step
up to left and become parallel to foliation, consistent with sinistral shear sense. Plane light. Scale bar is 250gin. (b) Outcrop
photograph of a and ~ feldspar grains on same flow plane in amphibolite gneiss, Parry Sound, Ontario. Tails on all
porphyroclast systems step up to left and become parallel to foliation, consistcnt with sinistral shcar sense. Coin diamctcr is
2.4 cm. Sec tcxt for discussion.

11
 C. SIMPSON a n d D . G . D E PAOR

R reference
axis ,~/~, (R,~)

/~'/ deformation
V path
Fig. 17. Plot of non-coaxial deformation path on the hyperbolic net.
Stretch and rotation factors are represented by pie segment.

E
Fig. 16. The hyperbolic net (see De Paor 1988 for a larger version). R is
axial ratio, shape factor E = (1/2) In (R).

(I

(~

)aseline

/ ......./
baseline

Fig. 18. Behavior of rigid objects plotted on the hyperbolic net. Thin straight lines are eigenvectors. Thin curved lines
represent trajectories, arrows indicate direction of rotation. Note back rotation on concave side of hyperbola (heavy line).
Stable end orientations marked by representative ellipse shapes. (a) Narrowing sub-simple shear, (b) Broadening sub-
simple shear. Squares represent undeformed states, and parallelograms deformed states.

12
 Strain and kinematic analysis in general shear zones 13

objects of known axial ratio, such as garnets that contain grains would indicate a high k/~, and b grains would
inclusions, with respect to the foliation around them it indicate a low/~/~) (Passchier & Simpson 1986). How-
may be possible to estimate Wn- This technique was ever, if the two types of grains are of the same mineral
successfully applied by Vissers (1989) to obtain Wn and and in the same rock then they cannot reflect great
the total strain from the rotation of pre-mylonitic gar- differences in the tempeature-controlled recrystalliza-
nets in a schist (see also Vissers 1987). tion rate. If they are on the same flow plane, or nearly
so, as in Fig. 15(b), then a large gradient in strain rate is
a and b grains also unlikely. One way to form grain associations like
the one in Fig. 15(b) is to start with grains of very
Porphyroclasts with recrystallized tails are extremely different axial ratio and/or orientation.
useful for sense of shear determination in high strain
zones (see Simpson & Schmid 1983, Lister & Snoke Analysis of strain and strain path using the hyperbolic net
1984, Passchier & Simpson 1986). The sense of shear is
given by the sense of 'stair-stepping' of the tails on either Building on the pioneering work of Ramberg (1975)
side of the porphyroclast. Tails that step up to the right and Ghosh & Ramberg (1976), we now demonstrate
give a dextral shear sense. Grains with wedge-shaped that the concepts of finite general shear and porphyro-
tails are o grains. Grains with very narrow tails that cross clast morphology provide a framework in which a great
the reference plane are b grains. The distinction is more variety of strain markers and kinematic indicators may
than pedantic. be understood. Porphyroclasts are indicators of both
Passchier & Simpson (1986) showed that the presence strain and flow, and the hyperbolic net of De Paor (1988)
of o or b tails depends on the rate of recrystallization/~ to can be used to analyze the vorticity and strain path as
shear strain rate ~. If R/~ is high, the flow of recrystal- well as the finite strain state.
lized material in the tails away from the porphyroclast is De Paor (1988) described the use of the hyperbolic net
continually supplemented by the production of new for the implementation of the Rf/~ technique of Ramsay
recrystallized grains, resulting in the wedge-shaped tails (1967), Dunnet (1969) and Dunnet & Siddans (1971),
of a grains. If R/~ is low, there is insufficient new the shape factor grid of Elliott (1970), and the 0-curve
recrystallized material added to the tail to form a wedge method of Lisle (1977, 1985). Strained objects are
and the narrow tail that results becomes dragged around assumed to be perfectly passive ellipses in all of these
with the rotating b-type host grain. The 6 grains are words. The hyperbolic net is a polar graph of ellipse
usually only found in high strain, ultramylonite zones. shape versus orientation (Fig. 16). Orientation ~ is
Passchier (1988) suggested that the degree of 'stair- measured around the periphery of the net and ellipse
stepping' of wedge-shaped recrystallized tails on ty-type shape is plotted radially outward from the origin. Dia-
porphyroclasts can be used to determine Wn. The metrically opposite points on this plot represent the
method is based on the assumption that o grains and same ellipse but there are two scales for describing
mylonite foliations give the instantaneous stretching ellipse shape: axial ratio R and shape factor E = 1/2
direction and the eigenvector for flow, respectively, In (R) (see De Paor 1988 for details). Note that a circle
regardless of the directions of the diplacement axes. (axial ratio = 1, shape factor = 0) plots at the origin.
Two sets of hyperbolae adorn the hyperbolic net. Those
Complex o-6 grains with vertical and horizontal asymptotes and diagonal
vertices are trajectories that trace the paths of ellipses
Some porphyroclast systems have more than one set undergoing strain in the direction of the net axis. Hyper-
of tails (e.g. Fig. 15a, and see Davidson et al. 1982, bolae with vertical and horizontal vertices and non-
Hanmer 1984). Passchier & Simpson (1986) reproduced orthogonal asymptotes are loci of equal incremental
these by starting with a b grain (low /~/~) and then strain and serve to determine the 50%-of-data curve
varying the R/~ parameters so that R increased with (Dunnet 1969), also known as the 0 = 45 ° curve (Lisle
respect to ~. Natural examples imply that: (a) the tem- 1977, 1985).
perature rose during a constant strain rate deformation; The hyperbolic net is used with a tracing overlay in a
(b) the shear zone remained at the same temperature manner analogous to a steronet. To represent a strain
during a decreasing strain rate deformation; or (c) the path, the cumulative finite strain ellipse is plotted for
shear zone maintained a constant temperature and successive points in time. A coaxial deformation in-
strain rate, but the grain itself was initially elliptical in volves no change in strain ellipse orientation and so is
cross-section, and therefore changed its rotation rate recorded by a straight path directed radially out from the
during deformation. origin. Increments of rigid rotation change the orien-
The existence of a, b and complex porphyroclast tation but not the shape of the strain ellipse and so plot
systems (Passchier & Simpson 1986) in the same outcrop on a circular arc. Shear in general follows a curved path
is not uncommon (e.g. Fig. 15b). Such occurrences (Fig. 17) from the initial point (1, q~) to the point
preclude Bell & Johnson's (1992) explanation of b grain representing the final state (R, q~'). The effects of an
tail geometry as the result of three separate, mutually increment of deformation upon a pre-existing ellipse
perpendicular orogenies. If all grains were identical in may be modelled by applying small increments of irrota-
mineralogy and were spherical to start with, then cr tional stretch and rigid rotation. The ellipse is moved
14 C. SrMPSON and D. G. DE PAOR

along a hyperbolic trajectory to simulate stretch and the

tracing overlay is rotated to simulate rigid rotation.

Analysis of vorticity using the hyperbolic net

The hyperbolic net can also be used to determine

another hyperbolic locus called the 'stable-end' curve
which gives a simple method for obtaining Wn from rigid
elliptical porphyroclasts in mylonites. The basis for this
method is the theoretical work of Jeffrey (1923), Ghosh
(a)
& Ramberg (1976), Freeman (1985) and Passchier
(1987) concerning the motion of ellipsoidal, perfectly
rigid particles in steady non-coaxial flow. Elongate
objects inclined downstream in a shear zone tend to
rotate backward if there is positive extension along the (b)
shear zone boundary, i.e. the shear zone is narrowing
(Fig. 18a), and elongate objects inclined upstream tend
to rotate backward if there is a contraction along the
boundary, i.e. the shear zone is broadening (Fig. 18b).
As illustrated in Fig. 18, lines in all orientations in the
acute angle between the eigenvectors rotate backwards,
but not all elliptical objects do so. A rigid sphere (R = 1)
is unaffected by the irrotational component of general
shear and responds only to the rotational component.
Therefore it will rotate in the direction of shear, albeit
more slowly than the shear strain rate (for simple shear it
would rotate at half that rate). Because all lines in the
acute field rotate backwards and spheres rotate for- Fig. 19. Behavior of deformable objects in (a) narrowing sub-simple
wards, there must be critically shaped and oriented shear, and (b) broadening sub-simple shear, depicted schematically on
the hyperbolic net. Details of trajectories (thin curved lines) depend
elliptical objects (R, q~) that do not rotate at all, and a on the ductility of porphyroclasts, e1 represents the infinitesimal
minimum axial ratio below which all objects rotate with stretching axis. Other features as in Fig. 18.
the flow.
In the Appendix we provide proof that all rigid
B = tanh (2E). (12b)
ellipses which do not rotate plot on a hyperbola asymp-
totic to the eigenvectors. Half of these are unstable Equation (11) is analogous to equation (A22) of De Paor
positions from which objects flee given a slight nudge. (1988), therefore the hyperbolic locus of stable end
The other half are the 'stable end' positions (those positions is given by fitting a curve from the hyperbolic
marked by a dot on Fig. 18) (Ghosh & Ramberg 1976, net asymptotically between the flow eigenvectors.
Freeman 1985, Passchier 1987). The orientation of a
stable ellipSe is a function of its axial ratio and of the Rotation of deformable objects
kinematic vorticity number Wn. The minimum axial
ratio, R m, for stable orientation plots at the apex of the Deformable ellipsoidal objects may possess internal
hyperbola and is related to Wn by vorticity which differs from that of the bulk flow. During
deformation, their long axes and axial ratios change and
Wn =- (Rm - ]/Rm)/(RmW1/Rm). (9) a succession of new material lines move through the
The orientation q~m of this individual is also related to immaterial axes of the ellipsoid. If objects are not
flOW, perfectly ellipsoidal, the migration of their axes may be
quite complex. We can never determine the vorticity of a
W n = cos (2q~m) (10a) single object, but nevertheless we can use a set of
= cos (v). (10b) deformable objects to determine the bulk flow vorticity.
The behavior in cross-section of deformable ellipsoidal
All other stably oriented ellipses plot along one limb of
grains is schematically illustrated in Fig. 19. Following
the hyperbola, whose formula is
Lisle (1985), we assume that the grain's rotation axis
B = Wn/cOs 2q~, (11) remains perpendicular to the flow direction (cf. Lister &
Price 1978). The obvious difference from Fig. 19 is that
where q~ is measured relative to q~,, and B is an alterna-
trajectories are no longer circular arcs. Elliptical cross-
tive description of ellipticity, related to the axial ratio,
sections become greater or smaller in axial ratio depend-
R, and Elliott shape factor, E, respectively by,
ing on the object's orientation relative to the fields of
B = (R - 1/R)/(R + 1/R) (12a) extensional and contractional infinitesmal strain. An
and object initially plotting on the concave side of the hyper-
 Strain and kinematic analysis in general shear zones 15

bola rotates backward while deforming until it reaches microstructures that fit this model include myrmekitic
its instantaneous stable end configuration. It then pro- feldspars (Simpson & Wintsch 1986), which are
ceeds to deform while rotating and follows a trajectory expected to be in stable end orientations, and fractured
that coincides with the hyperbola. Eventually, all back- feldspars which can form in a variety of orientations but
rotated objects become extremely stretched and which are most likely to fracture when their cleavage
oriented virtually parallel to the stable eigenvector, i.e. planes are aligned along directions of maximum shear
parallel to the foliation. Deformable objects that in- strain rate (Fig. 20).
itially plot on the convex side of the hyperbola display One problem with this approach is that there may be
complex trajectories. Most rotate forward, either spin- little data in the region around the unstable eigenvector
ning rapidly and pulsating in axial ratio if they were because all elongate objects rotate away from this line.
intially relatively fat, or slowing down and reaching the Therefore, just as in shear sense determination where
hyperbola if they were elongate. However, some objects correct interpretation depends on the statistical distri-
that plot near the hyperbola undergo a period of back- bution of grains, a large population of grains is needed,
rotation as their axes spin backwards and sweep through the maximum combination of methods to determine Wn
material lines. should be used, and the foliation should represent a
single, progressive deformation. The analysis should
Practical implementation include o and 6 or complex 6-0 type porphyroclasts, and
there must be relatively equant grains in addition to
Axial ratios R and orientations ~ of porphyroclasts highly elongate ones as the former give more accurate
are measured relative to the shear zone boundary, or results. Thin sections should be cut thicker than normal
relative to foliation in a narrowing shear zone in which to reveal the orientation of the porphyroclast in the third
shear strain is judged to be so high that foliation and dimension; individuals that are cut obliquely or that
boundary are virtually parallel. Axial ratio and orien- impinge upon their neighbors should be rejected (Pass-
tation data are plotted on the hyperbolic net, along with chier 1987). The key to successful application of this
shear sense, using symbols to distinguish back- and technique is the recognition in the rock of objects that:
forward-rotated o grains, 6 and complex 6-0 grains. Just (1) are rigid vs those that are deformable; (2) are in their
as with steronet data, the points when plotted are quite stable end positions vs 'immature individuals'; and (3)
independent of their location in the rock. The net is have rotated backward vs those (more common) that
rotated like a stereonet and a hyperbola (representing have rotated forward.
the locus of stable ends) is drawn asymptotic to the shear Narrowing vs broadening shear zones can be dis-
zone boundary direction so as to enclose the field of tinguished on the basis of their preserved microstruc-
back-rotated or o-type grains and then the hyperbola's tures only if the shear zone boundary orientation is
other asymptote is drawn. This second asymptote is known (Fig. 20). In the case of broadening zones, the
inferred to be the inclined eigenvector and the cosine of stable eigenvector and therefore the foliation, is main-
its inclination v to the flow plane is Wn (equation 1). tained at a significant angle to the zone boundary which
Takagi & Ito (1988) and Malaveille & Ritz (1989) is now the unstable eigenvector.
have described ellipsoidal porphyroclasts that are in-
clined downstream and have o tails attached to their Flow eddies and fabric disturbance
broad sides. They record backward rotation and plot
inside the hyperbola between the flow eigenvectors (Fig. The existence of stiff or relatively ductile grains in a
20). Objects such as elongate garnets that record back- flowing matrix leads to a complex, locally spinning flow
ward motion also plot in this field. 6 grains, which pattern. Figure 21(a) illustrates the streamlines of a
indicate continuous forward rotation, occupy the field flowing mass that includes a stiff grain, itself undergoing
outside of the hyperbola. Complex 0--6 grains also super-simple shear, in a relatively ductile matrix under-
indicate continuous forward motion, with formation of going sub-simple shear (the opposite case, Fig. 21b, is of
the o trails when the rotation of 6 grains slows down as mainly theoretical interest). The pattern is centered on
they approach the stable end orientation (Fig. 20). The the grain causing the disturbance. The important impli-
upstream-inclined mica 'fish' of Lister & Snoke (1984) cation of Fig. 21(a) is the disturbance of the ideal sub-
commonly have extremely thin and straight tails, or no simple shear flow pattern around the boundary of the
tails at all, consistent with a generally low/~/~. They may stiff grain. Drag on the grain's surface, combined with
have rotated forward without producing tails until they deflection of the flow lines, may lead to a situation in
approached their stable end orientation where rotation which smaller grains in a satellite position undergo a
rates would be slow enough to allow recrystallization to local rotational couple (Fig. 21a). Rotation may be
form long, straight tails. Ellipsoidal a grains that have opposite to that of the main vortex, even where grains
tails attached to their narrow ends and that are inclined are circular or near-circular in cross-section.
upstream may have had starting orientations close to
their final orientations. In addition, deformable Asymmetric boudins and sheath folds
upstream-inclined grains are subject to slight back-
rotation in the zone of complexity where the trajectories The distribution of asymmetric rigid or deformable
cross the stable eigenvector (Figs. 19 and 20). Other boudins and folds may also be controlled by eigenvector
16 C. SIMPSONand D. G. DE PAOR

(a) wn = .75

;/max

foliation ---- - sh arzono oun a

Yl / X ......

(b) wn =-.75

max... ...""

shear zone.boundary

.... "-'"--_.._

max

Fig. 20. Synoptic diagram of meso- and micro-scale structures in a dextral sub-simple shear zone with (a) Wn = 0.75
(narrowing) and (b) Wn = -0.75 (broadening). Cartoons represent various features in their approximate position and/or
orientation on the hyperbolic net. Hyperbolae are stable ellipse orientations as in Fig. 18. Solid lines are eigenvectors,
dotted lines are direction of maximum shear strain rate. See text for discussion.

orientations. If a compositionally stiffer band is parallel 22(f). If there is a planar compositional band parallel to
to the stable eigenvector (Figs. 22a-f), it may become the unstable eigenvector (Fig. 22g) then asymmetric
boudinaged. Subsequent rotation of boudin fragments folds may develop during shear (Fig. 22b), which may
may be in the same sense as the shearing for brittle rigid subsequently amplify into sheath folds.
boudinage or backward in the case of deformable frag-
ments (see Hanmer 1986). If a rigid layer (Fig. 22a) Strain partitioning
breaks into rigid blocks (Fig. 22b), the points rep-
resenting block shapes on the hyperbolic net are shifted Strain in shear zones is frequently partitioned into
towards the origin and thus enter the field of forward bands of intense shear and intervening lithons of less
rotation (Fig. 22c). Rotations cease when block orien- intense shear (e.g. Coward 1976, Simpson 1983, Bell
tations again plot on the locus of stable ends. A stiff but 1986). Bell & Johnson (1992) and Bell et al. (1992 and
more deformable layer (Fig. 22d) may break into self-references therein) have suggested that porphyro-
deformable fragments (Fig. 22e), which then back- blasts within such lithons do not rotate (but see Passchier
rotate (Fig. 22f) due to counterclockwise spin of their et al. 1992). Bell & Johnson (1992) have explained such
axes through material lines, as may be demonstrated by features as ~ grains by invoking a series of three different
drawing elliptical fragment outlines on a card deck and and orthogonal phases of deformation (their fig. 25), but
then shearing the cards. The corresponding trajectory this is refuted by the occurrence of o and 6 grains on the
on the hyperbolic net is illustrated by the arrows in Fig. same flow plane. Even in pure shear, elongate grains
 Strain and kinematic analysis in general shear zones 17

(3) where the shear zone is curved or has non-parallel

sides; or (4) where there is a fault between the wall rock
and the sheared rock. For a straight, parallel-sided shear
zone of constant cross-sectionial area, simple shear is the
only possible deformation regime if the wall rock re-
mains undeformed and attached (Ramsay & Graham
1970). There are, however, several tectonic scenarios
that provide the means for more general shear. Area
changes may be facilitated where shear zones rise to the
surface of the earth or where nappes spread, as often
occurs in the internal parts of orogens. In transpression
or transtension zones, material moves laterally into or
out of the profile plane. In the case of extreme transpres-
sion, the maximum principal stretch may be the axis of
the rotational component of deformation (Fig. 23), and
a mineral lineation could develop orthogonal to the
directions of tails of oblate grains (prolate grains are
only transiently stable in this case; Passchier 1987).
Broadening sub-simple shear zones may be dis-
tinguished by the relatively large angle between foliation
and shear zone boundary, even at very high strains, and
by the occurrence of dilatational structures such as
intrusive sheets parallel to the boundary (Escher et al.
1975). Both broadening and narrowing shear zones may
occur in both extensional and compressional tectonic
settings. It is important to consider zone orientations
relative to regional d6collements before making large-
scale inferences from the geometry of mesoscale general
shear deformation.

Fig. 21. (a) Disturbance of flow lines for sub-simple shear around a CONCLUSIONS
domain of super-simpleshear such as a deformableporphyroclastin a
narrowing zone. Note that trajectories represent particle paths as in
Fig. 2, and not ellipses as in Fig. 18. Flowdisturbance may cause back- The polar Mohr constructions presented here for
rotation of satellite grains even if they are circular in cross section. finite and infinitesimal strain, flow and acceleration, are
Eigenvectorslie in planes of shear becausethere are different displace-
ment rates along particle paths on either side (labelled 'faster' and ideally suited for the analysis of general shear in high
'slower'). (b) Flow lines within and around a soft grain undergoingsub strain zones. The kinematic vorticity number Wn is
simple shear in a matrix that undergoes super simple shear. readily obtainable from the angle between the eigen-
vectors, where one eigenvector is fixed to the flow plane
of the simple shear component, and the orientation of
must rotate towards the principal stretch axis by an the other depends on the pure shear contribution.
amount that depends on their axial ratio and initial Narrowing vs broadening sub-simple shear zones can be
orientation (Dunnet 1969, Ghosh & Ramberg 1976, distinguished using the relationship between foliation
Lisle 1977, 1985). In addition, three separate and ortho- and shear zone boundary. Super-simple shear is gener-
gonal orogenic events would cause shear zone bound- ally confined to the region in and around deformable
aries to be decidedly non-planar, and yet 6 grains are porphyroclasts.
most commonly foand in parallel-sided ultramylonite The hyperbolic net allows a simple and rapid analysis
zones. We see no reason to invoke a special cause for 6 of Wn using the stable end orientations of rigid and
grains in thick shear zones as opposed to narrow ones, deformable a, 6 and complex a-~ porphyroclasts. Other
nor do we find any inconsistencies between Bell & kinematic indicators such as myrmekitic feldspars, mica
Johnson's (1992) figs. 1-3 and 10, and a single event, fish, asymmetric boudinage and calcite lattice-preferred
general shear history. orientations are consistent with this approach. Fractures
in feldspar grains and shear bands may form along the
Tectonic implications of sub-simple shear directions of maximum shear strain rate and sub-
sequently rotate with progressive strain; if formed late in
The tectonic implications of sub-simple shear zones, the deformation history their orientations may allow an
whether narrowing or broadening must be carefully approximation of the general shear state.
considered. These shear zones may occur only: (1)
where the wall rocks to a high strain zone are able to Acknowledgements--The impetus for this work began with the joint
deform; (2) where there is an area change in the section; meeting of the Dutch and German Geological Societiesin Bochum,

15;14
18 C . SIMPSON a n d D . G . D E PAOR

(a) (d)

iiiii!i'ilil
iilii!iiii~ii~'
i ii!i!i',,iili
iii~i~~iii~,i'li'~iii~i,

t\\ I l\
ii~iiiii!iiiiiiii!~iiiiiiiiiiiiiiiiiilii~iiiiii!iiiii;i iiiii~i!

Fig. 22. Explanation of asymmetric boudinage and fold geometry in general shear zones. (a) Initial rigid layer.
(b) Boudinage into rigid fragments that rotate forward (c) into stable end orientations. (d) More deformable stiff layer
breaks into deformable boudins (e). (f) Deformable fragments may rotate backwards owing to spin of their axes. (g) Stiff
layer oriented parallel to unstable eigenvector. (h) Individual fold axial planes are sub-parallel to the shear zone boundary.

lcase time from the Division of Earth Sciences, U.S. National Science
Foundation.

REFERENCES

Allison, I. 1984. The pole of the Mohr diagram. J. Struct. Geol. 6,

331-334.
Behrmann, J. H. 1987. A precautionary note on shear bands as
kinematic indicators. J. Struct. Geol. 9, 659-666.
Bell, T. H. 1986. Foliation development and refraction in metamor-
phic rocks: reactivation of earlier foliations and decrenulation due
Fig. 23. Kinematic rotation axis may be parallel to S I in a zone of to shifting patterns of deformation partitioning. J. metamorph.
transpression. Geol. 3, 109-118.
Bell, T. H. & Johnson, S. E. 1992. Shear sense: a new approach that
resolves conflicts between criteria in metamorphic rocks. J. meta-
morph. Geol. 10, 99-124.
Bell, T. H., Johnson, S. E., Davis, B., Forde, A., Hayward, N. &
Wilkins, C. 1992. Porphyroblast inclusion-trail orientation data:
Germany, in 1990 and was further vitalized during preparations for the eppure non son girate. J. metarnorph. Geol. 10, 295-307.
1991 Geological Society of America Structural Geology and Tectonics Bobyarchick, A. R. 1986. The eigenvalues of steady state flow in Mohr
Division short course on Deformation and Kinematics in High Strain space. Tectonophysics 122, 35-51.
Zones presented by the authors in San Diego, California. The partici- Brace, W. F. 1961. Mohr construction in the analysis of large geologic
pants in the short course are thanked for their enthusiasm in putting strain. Bull. geol. Soc. Am. 72, 1059-1080.
these ideas to the test. Reviews by Chris Talbot and an anonymous Choukroune, P., Gapais, D. & Merle, O. 1987. Shear criteria and
reviewer were most helpful. C. Simpson acknowledges research re- structural asymmetry. J. Struct. Geol. 9, 525-530.
 Strain and kinematic analysis in general shear zones 19

Coward, M. P. 1976. Strain within ductile shear zones. Tectonophysics Passchier, C. W. 1986. Flow in natural shear zones--the consequences
34, 181-197. of spinning flow regimes. Earth Planet. Sci. Lett. 77, 70-80.
Curie, P. 1894. Sur la symetrie dans les ph6nom~nes physiques: Passchier, C. W. 1987. Stable positions;of rigid objects in non-coaxial
symetrie d'um champ 61ectrique et d'un champ magndtique. ]. Phys. flow--a study in vorticity analysis. ]. Struet. Geol. 9,679-690.
3, 393-415. Passchier, C. W. 1988. Analysis of deformation paths in shear zones.
Davidson, A., Culshaw, N. G. & Nadeau, L. 1982. A tectonometa- Geol. Rdsch. 77,308-318.
morphic framework for part of the Grenville Province, Parry Sound Passchier, C. W. & Simpson, C. 1986. Porphyroclast systems as
region. Ontario. Geol. Surv. Can. Pap. 82-1A, 175-190. kinematic indicators. J. Struct. Geol. 8, 831-843.
Dennis, A. & Secor, D. T. 1987. A model for the development of Passchier, C. W., Trouw, R. A. J., Zwart, H. J. & Vissers, R. L. M.
crenulations in shear zones with applications from the Southern 1992. Porphyroblast rotation: eppur si muove? J. metamorph. Geol.
Appalachian Piedmont. J. Struct. Geol. 9,809-817. 10, 283-294.
De Paor, D. G. 1983. Orthographic analysis of geologic structures--1. Platt, J. P. 1984. Secondary cleavages in ductile shear zones. J. Struct.
Deformation theory. J. Struct. Geol. 5,255-278. Geol. 6,439-442.
De Paor, D. G. 1988. Rf/$f strain analysis using an orientation net. J. Platt, J. P. & Vissers, R. L. M. 1980. Extensional structures in
Struct. Geol. 10, 323-333. anisotropic rocks. J. Struct. Geol. 2, 397-41/).
de Sitter, L. U. 1956. Structural Geology. McGraw-Hill, New York. Ramberg, H. 1975. Particle paths, displacement, and progressive
Dunnet, D. 1969. A technique of finite strain analysis using elliptical strain applicablc to rocks. Tectonophysics 28, 1-37.
particles. Tectonophysics 7, 117-136. Ramsay, J. G. 1967. Folding and Fracturing of Rocks'. McGraw-Hill,
Dunnet, D. & Siddans, A. W. B. 1971. Non-random sedimentary New York.
fabrics and their modification by strain. Tectonophysics 12,307-325. Ramsay, J. G. 1980. The crack-seal mechanism of rock deformation.
Elliott, D. 1970. Determination of finite strain and initial shape from Nature 284, 135-139.
deformed elliptical objects. Bull. geol. Soc. Am. 81, 2221-2236. Ramsay, R. G. & Graham, R. H. 1970. Strain variations in shear belts.
Elliott, D. 1972. Deformation paths in structural geology. Bull geol. Can. J. Earth Sci. 7,786-813.
Soc. Am. 83, 2621-2638, Ramsay, J. G. & Huber, M. I. 1987. The Techniques of Modern
Escher, A., Escher, J. C. & Watterson, J. 1975. The rcorientation of Structural Geology, Volume 2: Folds and Fractures. Academic
the Kangamuit dyke swarm, west Greenland. Can. J. Earth Sci. 12, Press, London.
158, 173. Ramsay, R. G. & Wood, D. S. 1973. The geometric effects of volume
Flinn, D. 1962. On folding during three dimensional progressive change during deformation processes. Tectonophysics, 16,263-277.
deformation. Q. J. geol. Soc. Lond. B 385-433. Rosenfeld, J. L. 1968. Garnet rotations due to the major Paleozoic
Freeman, B. 1985. The motion of rigid ellipsoidal particles in slow deformations in southeast Vermont. In: Studies of Appalachian
flows. Tectonophysics 113, 163-183. Geology: Northern and Maritime (edited by White, W. S., Zen, E.,
Gay, N. C. 1968. The motion of rigid particles embedded in a viscous Hadley, J. B. & Thompson, J. B). John Wiley, New York, 185-202.
fluid during pure shear deformation of the fluid. Tectonophysics 3, Sanderson, D. J. 1976. The superposition of compaction and plane
81-88. strain. Tectonophysics 30, 35-54.
Ghosh, S. K. & Ramberg, H. 1976. Reorientation of inclusions by Sanderson, D. J. & Marchini, W. R. D. 1984. Transpression. J. Struct.
combination of pure and simple shear. Tectonophysics 34, 1-70. Geol. 6,449-458.
Hanmer, S. 1984. Structure of the junction of three tectonic slices, Schley, H. M. 1973. Div, Grad, Curl, And All That: An Informal Text
Ontario Gneiss segment, Grenville Province. Geol. Surv. Can. Pap. on Vector Calculus. W. W. Norton & Co. Inc., New York.
84-1B, 109-120. Schmid, S. M. & Casey, M. 1986. Complete fabric analysis of some
Hanmer, S. 1986. Asymmetrical pull-aparts and foliation fish as commonly observed quartz c-axis patterns. Am. Geophys. Un.
kinematic indicators. J. Struct. Geol. 8, 111-122. Geophys. Monogr. 36,263-286.
Hanmer, S. & Passchier, C. W. 1991. Shear sense indicators: A Schoneveld, C. 1977. A study of some typical inclusion patterns in
review. Geol. Surv. Can, Pap. 90-117. strongly paracrystalline rotated garnets. Tectonophysics 39, 453-
Harker, A. 1995. On slaty cleavage and allied rock structures with 471.
special reference to the mechanical theories of their origin. Rep. Br. Simpson, C. 1983. Displacement and strain patterns from naturally
Ass. 55th Meeting, 1-40. occurring shear zone terminations. J. Struct. Geol. 5,497-506.
Harris, L. B. & Cobbold, P. R. 1985. Development of conjugate shear Simpson, C. & Schmid, S. M. 1983. An evaluation of criteria to
bands during bulk simple shearing. ]. Struct. Geol. 7, 37-44. detcrmine the sense of movement in sheared rocks. Bull. geol. Soc.
Jeffrey, G. B. 1923. The motion of ellipsoidal particles immersed in a Am. 94, 1281-1288.
viscous fluid. Proc. R. Soc. Lond. AI02, 161-179. Simpson, C. & Wintsch, R. P. 1986. Evidence for deformation-
Lisle, R. J. 1977. Estimation of tectonic strain ratio from the mean induced K-feldspar replacement by myrmekite. J. metamorph.
shape of deformed elliptical markers. Geologie Mijnb. 56, 140-144, Geol. 7,261-275.
Lisle, R. J. 1985. Geological Strain Analysis: A Manual for the Rf/q~ Sorby, H. C. 1855. On slaty cleavage as exhibited in the Devonian
Technique. Pergamon. London. limestones of Devonshire. Phil. Mag. 11, 20-37.
Lister, G. S. & Price, G. P. 1978. Fabric development in a quartz- Takagi, H. & Ito, M. 1988. The use of asymmetric pressure shadows in
feldspar mylonite. Tectonophysics 49, 37-78. mylonites to determine the sense of shear. J. Struct. Geol. 10, 347-
Lister, G. S. & Snoke, A. W. 1984. S-C mylonites. J. Struct. Geol. 6, 360.
617-638. Talbot, C. J., 197(I. The minimum strain ellipsoid using quartz veins,
Lister, G. S. & Williams, P. F. 1983. The partitioning of deformation Tectonophysics 9, 47-76.
in flowing rock masses. Tectonophysics 92, 1-33. Talbot, C. J. & Jackson, M. A. 1987. Internal kinematics of salt
Malavieille, J. & Ritz, J. F. 1989. Mylonitic deformation of evaporites diapirs. Bull. Am. Ass. Petrol. Geol. 71, 1068-1098.
in d6collements: examples from the Southern Alps, France. J. Truesdell, C. 1954. The Kinematics of Vorticio'. Indiana University
Struct. Geol. 11,583-590. Press, Bloomington.
March, A. 1932. Mathematische theorie der Regclung nach der Urai, J. L., Williams, P. F. & van Roermund, H. L. M. 1991.
Korngestalt. Z. Kristallogr. 81,285-297. Kinematics of crystal growth in syntectonic fibrous veins. J. Struct,
Matthews, P. E., Bond, R. A. B. & Van den Berg, J. J. 1974. An Geol. 13, 823-836.
algebraic method of strain analysis using elliptical markers. Tec- Visscrs, R. L. M. 1987. The effect of foliation orientation on the
tonophysics 24, 31-67. inferred rotation axes and rotation angles of rotated porphyroblasts.
Means, W. D. 1982. An unfamiliar Mohr circle construction for finite Tectonophysics 139, 275-283.
strain. Tectonophysics 80, T1-T6. Vissers, R. L. M. 1989. Asymmetric c-axis fabrics and flow vorticity: a
Means, W. D. 1983. Application of the Mohr-circle construction to study using rotated garnets. J. Struct. Geol. 11,231-244.
problems of inhomogeneous deformation. J. Struct. Geol. 5,279- Weijermars, R. 1991. The role of stress in ductile deformation. J.
286. Struct. Geol. 13, 1061-1078.
Means, W. D., Hobbs, B. E., Lister G. S. & Williams, P. F. 1980. Weijermars, R. & Rondeel, H. E. 1984. Shear band foliation as
Vorticity and non-coaxiality in progressive deformations. J. Struct. indicator of sense of shear: field observations in central Spain.
Geol. 2,371-378. Geology 12,603~06.
Nadai, A. 1963. Theory of Flow and Fracture of Solids. McGraw-Hill, Wenk, H.-R., Takeshita, T., Bechler, E., Erskine, B. G. & Matthies,
New York. S. 1987. Pure shear and simple shear calcite textures. Comparison of
Owens, W. H. 1973. Strain modification of angular density distri- experimental, theoretical and natural data. J. Struct. Geol. 9,731-
butions. Tectonophysics 16, 24%261. 745.
20 C. SIMPSONand D. G. DE PAOR

Wettstein, A. 1886. Uber die Fischfauna des Terti~iren Glarner Translating equation (A20) of De Paor (1988) using the identity
Schiefers. Schweiz. Paliiont. Ges. Abh. 13, 1-101.
B = tanh (2E), (A2)
we propose that the polar equation of the hyperbola in Fig. A1 is,
APPENDIX B = Wa/cos 2q), (ll)
where the kinematic vorticity number W, is the value of B at the apex
Passchier (1987) gives the following formula for the stable end and also the cosine of the angle between the asymptotes,
position of a rigid elongate porphyroclast, in a general shear zone in
which the pure shear component acts to narrow the zone: W~ = cos (v), (A3)
r/= (1/2) sin -~ { ( W , / B ) [ ~ / ( 1 - W 2) +_ X/(B 2 - W2)]}, (A1) The following derivation demonstrates that equation (11) of this paper
is equivalent to equation (A1) above and provides an easier way to
where B is the shape factor, and q is the orientation of the long axis of calculate W,. Note the different sense in which angles are measured in
the grain (Fig. AlL the paper (Fig. A1) vs Passchier (1987).
From equation (11), substituting for 2¢,
Wn/B = cos (v - 2r/) (A4)
= cos v cos 2q + sin v sin 2r/ (A5)
= W, cos 2q + sin v sin 2r/. (A6)
Rearranging, and then squaring both sides,
Wn cos 2r/= W . / B - sin v sin 2q (A7)
~ W n W2 (1 - sin e 2r/) = W2/B 2 - ( 2 W , / B ) sin v sin 2r/+ (1 - W2) sin e 2r/
(AS)
0 = W~ ( ( 1 / B ) - B ) - 2 sin v sin 2r/+ (B/Wn) sin22r/.
(A9)
Fig. A1. Demonstration that stable end orientations plot on a curve
from the hyperbolic net by showing that equation (11) of this paper is Solving this quadratic,
equivalent to Appendix equation (A1). v is the angle between asymp- sin 2r/= [2 sin v + # ( 4 s i n 2 v - 4 ( 1 - Be))]/(2B/Wn) (A10)
totes, W, is apical value of radius vector B which is related to ellipse
axial ratio R and shape factor E by equations (12a) and (12b). ¢ is the = (Wn/B)[~/(1-W2)+~/{(1-W.2)-(1-BZ)}] (All)
angular measure used in this paper. Passchier (1987) uses the alterna- or
tive angle r/. q = (I/2) s i n - ' { ( W . / B ) [ ~ / ( 1 - W 2) ++-\ / ( B z - we)]}. (A1)