Magnetic fabrics and superimposed deformations: example of Dalradian rocks from the southwest Highlands of Scotland

Physics of the Earth and Planetary Interiors, 40 (1985) 187—200 Elsevier Science Publishers By., Amsterdam — Printed in The Netherlands

187

Magnetic fabrics and superimposed deformations: example of Dairadian rocks from the southwest Highlands of Scotland Bernard Henry Laboratoire de Géomagnêtisme, 4 avenue de Neptune, 94107 Saint-Maur Cedex (France)

Henry, B., 1985. Magne ibrics and superimposed deformations: example of Dalradian rocks from the southwest Highlands of Scotland. Phys. Earth Planet. Inter., 40: 187—200. Analysis of anisotropy differences in specimens close in the same sample permits the separation of anisotropies due to ferrimagnetic minerals and their dia-paramagnetic matrix. This method, previously described, has been applied to rocks, affected by superimposed deformations, from the Dalradian of Scotland. It appears that ferrimagnetics anisotropy is often found in relation with the secondary deformation, even though matrix anisotropy is related to the primary deformation. This last anisotropy has here a dominant effect on the orientation of the resultant fabric of the whole rock.

I. Introduction It is now well established that magnetic susceptibility anisotropy is a useful method for qualitative structural analysis (Graham, 1954; Daly, 1970; Hrouda, 1982). Not only has the coincidence of orientation of visible structures (foliation, lineation) with their magnetic equivalents (plane and axis of maximum susceptibility) been shown in numerous examples, but also barely or non-visible strains were revealed with this method (Henry, 1973, 1974, 1980). For some years, researches have been more orientated to the quantitative relations between tensors of susceptibility and finite strain (Kneen, 1976; Wood et al., 1976; Hrouda, 1979; Rathore, 1979, 1980; Henry, 1980; Kligfield et al., 1981, 1982). The aim of these works was obviously the prospect of the use of susceptibility anisotropy data for a determination, at least relative, of finite strain. However, it was found that no universal correlation coefficient exists between tensors of susceptibility and finite strain (Rathore and Henry, 1982). This is not surprising since the magnetic susceptibility anisotropy of a rock depends on the anisotropy of each mineralogical family which 0031.9201/85/$03.30

© 1985 Elsevier Science Publishers B.V.

composes it (Daly, 1967; Coward and Whalley, 1979). The prospect of a finite strain calibration lies therefore in the separation of component anisotropies. Now, it was shown (Henry, 1983; Daly and Henry, 1983) that the analysis of anisotropy variations in specimens close in the same sample allows the separation of anisotropies due to ferrimagnetics and their dia-paramagnetic matrix. A first application (Henry and Daly, 1983) has been carried out on a limited collection of rocks, which were affected by a single deformation. The present work studies a large collection of rocks which underwent superimposed deformations (Dairadian in the southwest Highlands of Scotland), and we examine the problem of the relations between component anisotropies and the different superimposed deformations.

2. Method for the separation of anisotropies due to the two components of a rock A rock submitted to a field acquires an induced magnetization J linearly proportional in low field to the intensity H of the applied field. The susceptibility K, which is the ratio of the induced magne-

188

tization to the field (K J/H), corresponds to a second rank symmetrical tensor. For the susceptibility values, a rock is formed of two very different components, the ferrimagnetics of high susceptibility and their matrix of low diamagnetic and (or) paramagnetic susceptibility. Its composi-

same sample, and it is Just this variation of respective percentages of these two components which is the basis of the method used for the separation of component anisotropies. In an orthonorm reference (Ox, Oy, Oz), each of the K11 terms (i, J x, y or z) of the susceptibil-

tion varies from one specimen to another in a

ity tensor of a homogeneous strained rock may be

=

=

K13 ii

U

20

00

K tOO ~

200 • —

U.

~

Fig. 1. K,~ values versus k~values from different specimens of sample 31, with associated regression lines. For symmetrical terms, scale of K,~is five times that of R~K,~and K are in 10-6 SI.

189

written (for more details, see Henry and Daly, 1983)

tion, but the method of linear regression itself allows proof of such a homogeneity if a perfect

K

linear relation exists between K,1 and K. In fact, inhomogeneous strain in a sample gives differences from one specimen to another in anisotropies of the matrix and of ferrimagnetics; different relations (A,1K + B,1) thus correspond to each specimen. In practice, on a diagram (K,1, K), this gives scattering of the points for each specimen around the straight line associated with the mean relation (A~.K+ ~); the line obtained by

=

A K+ B

‘~‘

~j

(1)

‘~

where A,~and B,~are two constants related to the susceptibilities of the two components, and where K is the mean susceptibility of the rock. The susceptibility tensor may be written A K+ B A K+B A i~ +B xx

—

K +~

xy —

xx

~

xz

XY

K +~

~

A~K+B~~ AZYK+BZY

Xz

—

K+~

A2~K+B~~

or, writing1 1K in1 matricial [A iK+FB (2) ‘J.I 1 ‘JJ 1 ~ The K value depending on the proportion of the two rock components, it clearly appears with this relation that anisotropy is also related to this proportion. The relation (2), moreover, shows that the knowledge of the mean susceptibilities F of pure ferrimagnetics (in this case, F K and F, 1 K,1) and A? of the pure matrix allows the entire determination of the susceptibility tensors =

regression in this case is generally very near this “mean” line, if the total variation of K in the sample is for important higher K,1 fluctuation a same (much K value) and ifthan we have numerous specimens. So, this method of linear regression is, on the one hand, a simple and rapid test to show the homogeneity of strain at the sample scale, a test which cannot be made by simple observation. It allows one, on the other hand, to obtain significant results of the mean anisotropy in the case of

=

=

offerrimagnetics and of the matrix

1 1

1

Fl ‘Jj

F

1 —

1

IA.IF+IB,.I 11 L

=

1

/

\

~3)

inhomogeneous strain. Finally, we may note that the relation (2) using the mean susceptibility value of the sample allows us to determine the susceptibility mean tensor for the whole rock.

[M,~11A111 M+ [B,11 (4) =

However, in the different terms of the susceptibility tensor of ferrimagnetics, B,1 is generally negligible compared to the products A,1 F; the diagonalisation of [A,J] thus allows the determination of principal directions, the principal values being defined to a nearest factor F. In practice (Henry, 1983), from the experimental results, the determination of the six (symmetrical tensor) pairs of coefficients (A,1, B,1) can be obtained by linear regression from the pairs of values (K,1, K) measured on the different specimens from the same sample (Fig. 1). On the basis of this method, one assumes that finite strain is homogeneous in the sample; in this case, matrix anisotropy as ferrimagnetic anisotropy are quite similar in all specimens, and only the proportion of the components varies effectively from one specimen to another. Such a homogeneity of finite strain cannot be strictly proved by simple observa-

3. Application to the Dairadian rocks from the southwest Highlands of Scotland The Dalradian series in the southwest Highlands is a lithologically diverse sequence of shales, limestones, dolomites, sandstones, breccia, lavas,... from the upper Precambrian to Upper Cambrian (possibly Lower Ordovician—Downie et al., 1971). It was first deformed during the Caledonian orogeny, at the end of the Cambrian to the beginning of the Ordovician period, then it underwent various deformations. The first period of deformation was responsible for the major folding, with the formation of the Tay nappe (Aberfoyle and Ardrishaig anticlines) and the Loch Awe synclinorium. Finite strain in these rocks is clearly related to the slaty cleavages and stretching lineation from this first period (Borradaile, 1979). Subsequent periods correspond to the secondary

190 TABLE I Values of parameters P

1, P2, P3 and F for whole rock (R) and for ferrimagnetics (F)

Site

Sample

A

I

A

2

A

3

A

4

A

5

B

20

B

21

B

22

B

23

C

7

C

8

D

6

D

9

E

24

E

25

E

26

E

27

E

28

F

10

F

11

F

12

G

29

G

30

G

31

P1

P2

P3

E

R F R F R F R F R F

1.0588 1.0366 1.0665 1.0934 1.0345 1.0163 1.0521 1.0987 1.0506 1.1257

1.0738 1.0604 1.1196 1.1482 1.1862 1.1488 1.1983 1.3929 1.2293 1.2559

1.0141 1.0230 1.0498 1.0501 1.1466 1.1303 1.1390 1.2678 1.1701 1.1156

0.9578 0.9869 0.9844 0.9604 1.1085 1.1121 1.0826 1.1540 1.1137 0.9911

R F R F R F R F

1.0249 1.0761 1.0229 1.1307 1.0298 1.0137 1.0197 1.5843

1.0561 1.1060 1.3204 1.9526 1.0760 1.1351 1.1606 6.3956

1.0305 1.0278 1.2907 1.7269 1.0449 1.1198 1.1382 4.0370 a

1.0056 0.9551 1.2616 1.5273 1.0147 1.1047 1.1163 2.5482

R F R F

1.0157 1.3150 1.1788 1.2585

1.2513 2.0771 1.9143 2.2477

1.2320 1.5796 1.6160 1.7859

1.2130 1.2013 1.3712 1.4190

R F R F

1.0255 1.1697 1.0990 1.0791

1.1191 2.6273 1.5629 1.5705

1.0915 2.2461 1.4227 1.4553

1.0649 1.9202 1.2960 1.3486

R F R F R F R F R F

1.1174 1.2147 1.0082 1.2562 1.1976 1.1965 1.0695 1.2448 1.0330 1.0432

1.2817 1.6038 1.2001 1.6239 1.4739 1.3546 1.1916 1.6378 1.1250 1.2378

1.1468 1.3203 1.1904 1.2927 1.2314 1.1321 1.1143 1.3158 1.0891 1.1865

1.0263 1.0870 1.1807 1.0290 1.0404 0.9462 1.0421 1.0570 1.0543 1.1374

R F R F R F

1.0083 1.0179 1.0091 1.3339 1.0044 1.0302

1.1591 1.1742 1.0490 1.7738 1.0219 1.0703

1.1496 1.1535 1.0395 1.3298 1.0174 1.0389

1.1401 1.1333 1.0301 0.9969 1.0129 1.0084

R F R F R F

1.0746 1.0911 1.0389 1.0465 1.0284 1.0118

1.1798 1.2192 1.1693 1.2347 1.1360 1.1342

1.0980 1.1175 1.1257 1.1798 1.1046 1.1210

1.0221 1.0242 1.0838 1.1274 1.0741 1.1079

a

a

a

191 TABLE I (continued) Site

Sample

H

13

H

14

H

15

H

16

I

17

I

18

I

19

a

P

P2

P3

F

R F R F R F R F

1.0126 1.0162 1.0205 1.0740 1.0117 1.0476 1.0259 1.1959

1

1.0687 1.3540 1.1658 1.5131 1.1122 1.2290 1.0994 1.2851

1.0554 1.3324 1.1424 1.4089 1.0993 1.1732 1.0717 1.0746

1.0422 1.3111 1.1194 1.3119 1.0865 1.1199 1.0446 0.8985

R F R F R F

1.0191 1.0357 1.0467 1.0360 1.0298 1.0974

1.0764 1.1761 1.0749 1.1212 1.1290 1.1925

1.0562 1.1356 1.0269 1.0822 1.0964 1.0867

1.0365 1.0964 0.9811 1.0446 1.0648 0.9903

Very unprecise value, due to the weak difference between mean susceptibility of the different specimens of the sample.

TABLE II Mean susceptibility K (in 10-6 SI), and absolute value of the coefficient r of linear regression (after rotation of 45°of the reference axes for symmetrical terms) for the six regressions of each sample (with the number N of specimens for each sample) Site

Sample

N

A A A A A B B B B C C D D E E E E E F F F G G G H H H H I I I

1 2 3 4 5 20 21 22 23 7 8 6 9 24 25 26 27 28 10 11 12 29 30 31 13 14 15 16 17 18 19

7 9 7 6 6 9 3 8 6 6 7 5 7 12 6 7 8 10 7 6 11 9 12 9 5 6 6 7 7 8 11

K 80 76 476 280 216 60 13 376 154 245 609 354 1268 1409 255 25 428 151 270 497 714 17537 165 211 690 521 535 430 80 46707 771

rxx

~

r,~

rxy

~

1 1 0.999 0.953 0.923 1 0.964 0.990 0.872 0.995 1 0.921 1 0.999 0.987 0.996 0.986 0.997 0.999 0.968 0.999 1 0.999 0.997 0.999 0.999 0.998 0.956 1 0.999 0.998

0.999 1 0.999 0.991 0.993 1 0.980 0.994 0.929 0.988 0.999 0.992 1 0.995 0.985 0.999 0.873 0.996 1 0.994 0.999 1 0.998 0.998 0.999 0.998 0.998 0.982 0.999 0.999 0.988

0.999 1 1 0.971 0.961 1 0.997 0.994 0.991 0.983 0.998 0.823 1 0.997 0.996 0.998 0.901 0.998 1 0.988 0.999 1 1 1 0.994 0.993 0.997 0.981 0.999 0.999 0.990

0.999 0.999 0.998 0.980 0.994 0.999 0.992 0.987 0.772 0.976 0.998 0.774 0.998 0.997 0.912 0.993 0.935 0.999 1 0.954 0.999 0.998 0.997 1 0.971 0.964 0.997 0.952 0.999 0.998 0.996

0.999 0.999 0.994 0.956 0.972 1 0.922 0.993 0.426 0.897 1 0.548 0.996 0.993 0.936 0.995 0.971 0.994 0.995 0.979 0.998 0.998 1 0.999 0.974 0.917 0.999 0.979 0.999 0.996 0.982

0.999 0.999 0.995 0.872 0.801 0.999 0.837 0.979 0.844 0.682 0.960 0.843 0.998 0.997 0.941 0.992 0.942 0.996 0.999 0.768 0.998 0.999 0.998 0.999 0.945 0.994 0.988 0.978 0.999 0.999 0.997

192

deformations, in particular with the formation of Tarbert and Ben Ledi folds and locally the development of southeast dipping crenulation cleavage, This cleavage has NE—SW intersections with primary cleavages, intersections which correspond to secondary fold axes. It is particularly evident near the Tarbert and Ben Ledi folds and in the northern part of the area studied. Figure 2 shows the localizations of the nine sites (A—I), in which 31 samples were collected; orienta-

deformation (particularly near the Tarbert fold). Principal values used by Rathore and Henry (1982) were measured on Digico CRAD; it has been shown since that in this instrument an error exists in the computing procedure of principal values (Hrouda et a!., 1983). Thus, these last values have been recalculated for the Dalradian from Scotland, and Table I gives the corrected values of parameters P1 K1/K~,P2 K1/K3, P3 K2/K3 and =

tion of visible structures (Borradaile, 1973; Roberts, 1974) and magnetic fabric of whole rock were analysed in detail in a previous work (Rathore and Henry, 1982); principal susceptibility axes coincide most often with their structural equivalents of the primary deformation, but for many samples they appear in relation with secondary

=

=

2

E (K2) /K1 K3 where K1, K2 and K3 are the principal susceptibilities (K1 ~ K~~K3). These parameters allow a direct comparison with the classical equivalent parameters for strain (Flinn, 1962), and Rathore (1979) attempted to establish an empirical relation =

N

_

.1’

‘.~H

/1

/

/

.11

/

/16

/

.12F

_____

/ ~~v’

~)1L~

/ / ,“1l.LI’t ~sIc)r19 r( ~

/

18 ~l7

.1

Jj~)

/ L,~ ~•)/ //J

(ii

/

~/

/

/

/ C

/ /

/

‘V

c

IC,

/~

‘V 4.

~

/

t/

/

10 km

•1o

/f If I•6

/ /1/If

“

/

/

/

-

b”

/ ,./ //,-‘~,

/ /7

cx / 1

Ii)’

/

-

,

/

/

/

/

~

/

~

~

/

$27 28

/

,.26E

/

//

/

~

~

~

‘\ ~

/

‘~\

~ .-

\,..‘

23../c~)

Fig. 2. Map of sampling sites (A_I), with sample numbers. Primary structures: Aberfoyle (a) and Ardrishaig (b) anticlines, and Loch Awe synclinorium (c_c). Secondary Structures: Ben Ledi (d) and Tarbert (e) folds, and Highlands boundary fault (f).

193

between corresponding axial ratios in ellipsoids of susceptibility and finite strain. In this lithologically diverse series of the Dairadian, the relation recalculated from values of Table I (loglO, P1 =

0.049 loglO (X/ Y) where X and Y are, respectively, the principal piaximum and intermediate values of finite strain) is much less close than in more homogeneous rocks previously studied (Rathore, 1979, 1980). This result is not surprising, since the proportion of ferrimagnetics and matrix is the most variable in the Dalradian, among the different studied series. For each of the 245 specimens derived from the 31 Dairadian samples, the six terms of the susceptibility tensor were calculated in a same reference (Ox magnetic North, Oy, Oz descending vertical). For each sample, six linear regressions were made from the six series of pairs of values (K,~,K) to obtain the corresponding pairs of coefficients (A,1, B,1). The values of the coefficient r of linear regression (Table II) are > 0.99 for 67% of the regressions, and > 0.8 for 97% of the regressions. Thus, the results obtained with this method are here very significant, and, in these sites, the strain is most often very homogeneous_at the sample scale. Without determination of F values, principal directions and axial ratios for ferrimagnetics were obtained by diagonalisation of [A,1]. 3.1. Comparison of magneticfabrics orientation Principal axes and parameters are denoted by the letters R, F and M, respectively, for whole rock, ferrimagnetics and matrix. Orientation (determined for a mean tensor within each sample, and with Fisher s analysis from these sample results within each site) of principal axes of the resultant anisotropy R and of the component anisotropy F is different (Fig. 3); thus, axes of the two component anisotropies F and M also have different orientations. The angular difference between R and F axes has, in seven out of the nine sites, very similar values for minimum and maximum axes. In each site, for each of the four kinds of axes (R maximum, F maximum, R minimum and F minimum), the mean value of angle A separating two same type axes from two different samples has

been calculated (Fig. 4). One could expect a better grouping of axes for the component anisotropy F than for the resulting anisotropy R (AR > A F); but that is true only in half of the cases. In several sites, the A values for minimum and maximum axes are very similar (for the F axes only, or for the F and R axes). These observations show a homogeneity in the difference of orientation between axes R and F, as in the axes scattering within each site; thus, orientation changes of the magnetic fabrics from one specimen to another are coherent. Therefore these changes are not due to a lack of precision in the determination of axis orientation, but are connected with local structures in the rocks. 3.2. Comparison with visible fabric Table III shows that F axes are more related to secondary deformation than R axes. These relations are strongest in zones where secondary deformations are most visible (neighbouring the Tarbert fold and northern zone: sites B, D, F, H and I). TABLE Ill

.

.

.

Relations between magnetic fabrics and deformation: number of axes having the same orientation (magnetic axis within the grouping of visible structures observed in the site) as the primary (Prim. def.) and the secondary (Sec. def.) equivalent structures (pole of cleavage for minimum axis, and lineation for maximum axis), or orientation different from the two previous groups (Diff.)

Prim. def. Sec. def. Diff

Minimum R

Maximum R

Minimum F

Maximum F

30 1 0

22 3 6

17 5 9

17 11 3

_______________________________________________

R axes, although they represent only the resultant anisotropy, mostly coincide with the primary structures. The simple fact that they are distinct from F axes already shows that, contrary to the useful views admitted by numerous authors, matrix anisotropy cannot be considered as negligible compared to ferrimagnetic anisotropy. Now, we attempt to specify the part of each component anisotropy in the orientation of the resultant ani-

194

sotropy. We saw previously that, in different samples, F axes are not in relation with the primary deformation, although R axes are connected with this deformation. Such an arrangement in several samples cannot be due to an accident; it seems to imply that the other component anisotropy, that of matrix, is still associated with primary structures. For five samples, chosen as a test, all the possible orientations of principal axes have been determined using relation (2) for K values between mean susceptibility of the sample and 1110-6 SI

(susceptibility of quartz, which is the most diamagnetic mineral in these rocks); in fact, these values necessarily enclose the A? value, and so the orientation of the M axes is included within the group. It is observed, that for values between the mean susceptibility of the sample and relatively low K values (probably lower than the true A? value, judging from the mineralogy of these rocks), orientation of principal axes coincides with the equivalent structures of the primary deformation. Thus, matrix anisotropy seems to be in relation with this primary deformation.

—

N

N

S

0

0 •

S

0

0

o~

~

‘.36

+0

V

V V VEr

V

V V

N

N

V

Uv ©~19

•.15 SD •0

0

V

/

~

V

+

oQ~

V

2%

195 N

N

(~~) ~ —i—

VVV

V

00 OOS S •1O S~

V

54\.

V

V

—(—-\

N

0SkO. V

V

N V

~27

v

V

~0 V V V

V

V

110 V S S

©

0

V

S

V

N

9

0

0

!34

0 V

•

V

S

S

V

•

A V

31,’

I V

,;

Fig. 3. Maximum (triangles) and minimum (circles) susceptibility axes for whole rock R (full symbols) and for ferrimagnetics F (open symbols) in the nine sites (A_I). Each mean direction of the four groups of axes (R maximum, R minimum, F maximum and F minimum) for each site is indicated by a bigger symbol, with the angle measured in degrees between same name (maximum or minimum) mean direction R and F. Stereographic projection in the lower hemisphere.

196

AR 80

60

D

40

H,

F

20

G

AF 0

20

40

60

80

Fig. 4. Mean values in degrees of the angle between two maximum axes (triangles) and two minimum axes (circles) for whole rock (AR) versus those for ferrimagnetics (A F). N

N

/I

12tJ \~3925

/

L7

V /

Fig. 5. All the possible orientations of maximum (triangles) and minimum (circles) axes for mean susceptibility values (in 10-6 51) between 59 (mean susceptibility of the sample) and —11, for sample 20 (stereographic projection in the lower hemisphere). Primary (lv) and secondary (1,) lineations, and primary (sr) and secondary (se) cleavages poles, after Roberts, 1974.

100

h~

197

(AS 23)

U U

U

•

U

50

• •

/

U U

U U

U

U

•

•/

.“/

•.~/ •

10

•/

// •

U

/

io

0

~o

—

N

10

I -tO

0

10

50

.®fl

. lOO

fl~ th~—h~t

Fig. 6. (a) Anisotropy percentage values h’F (ferrimagnetics) versus h’R (whole rock). (b) Number N of samples as a function of (h’F —

198

=

However, we have also previously remarked that three samples have R axes coinciding with the secondary structures. It was, therefore, interesting to find out whether M axes in these three samples are also related to the primary deformation. With this object, all the possible onentations of M axes have been determined using the method described in the preceding paragraph. For two samples (from site F) of which the maximum R axis coincides with the secondary

rock; it should be 4.8 for a matrix if A? 38 10-6 SI and 7.1 if A?= 25 10-6 SI. In this sample, matrix is less anisotropic than ferrimagnetics. Figure 7 shows values of the shape parameter

lineation, the maximum axis remains different from the primary stretching lineation whatever A? value; the M axis should be connected with the secondary lineation, except if the M value is low. For sample 20 (from site B), maximum and minimum R axes are in relation with the secondary deformation (Rathore and Henry, 1982). Figure 5 shows that for K values (in 10—6 SI) between 25 and 38, the fabric of this sample is connected with the primary deformation. The coincidence for these values of both minimum and maximum axes with, respectively, the slaty cleavages pole and the stretching lineation is probably not a simple accident. The mean susceptibility of matrix A?, in this sample made up fundamentally of quartz and some sheetsilicates, is probably of this order. Sample 20 has in this case, as does the great majority of the others, M axes related to the primary deformation. The majority of M and R axes seem to have strong relations with the primary deformation. Matrix has thus a dominant effect relative to ferrimagnetics for the orientation of the resultant fabric R.

it gives significant results for all susceptibility val-

3.3. Comparison of quantitative data

0.5

—

) (

—

)

—

for whole rock (fR) as a function of those (fF) for ferrimagnetics; this parameter can vary from 0 (foliation without lineation) to 1 (lineation only), the value 0.5 corresponding to K2 (K1 + K3)/2;

=

N

— — —

—

____________

— —

-0,1

0

—

—

_________

0,5

0,1

1 f F

@J •

•

Figure 6 shows values of corrected (Jelinek, 1981) anisotropy percentage

U

—

//

//

• U

/

/

~“~“

• •

h’=100(1K1—KI-4-1K2—KI+1K3—K!)/K for whole rock (ha) as a function of that (h~)for ferrimagnetics; h’F is often higher than h~.Values of another parameter of anisotropy intensity, P2 (Table I), show the same difference between ferrimagnetics and whole rock. In sample 20, for which the order of magnitude of M is known, h’ values are 11.8 for ferrimagnetics and 5.6 for whole

U 11/

/T~ .

/ 0

1

Fig. 7. (a) Shape parameter values JF (ferrimagnetics) versus JR (whole rock). (b) Number N of samples as a function of (/F

—

JR)

199

ues. Shape differences between ellipsoids R and F’ are relatively strong; for the major part of the samples, the ferrimagnetics ellipsoid is more prolate than the whole rock ellipsoid. Values of the shape parameter E (Table I) reach the same conclusions. However, we cannot deduce from these data that the matnx ellipsoid is more oblate than the whole rock ellipsoid, the sum of two prolate ellipsoids being in some cases an oblate ellipsoid, Moreover, for sample 20 the f value, of 0.74 for ferrimagnetics and 0.46 for whole rock, should be 0.15 if A? 38 10-6 51 and 0.85 if A? 25 10-6 SI. Further, because of the non-coincidence of principal axes of finite strain and of ferrimagnetics susceptibility, the quantitative data of these two fabrics cannot be simply associated. In these Dairadian rocks, one cannot hope to establish a finite strain calibration with only the magnetic fabric of the ferrimagnetics.

=

=

4. Conclusion Susceptibility anisotropy of ferrimagnetic minerals in these Dalradian samples is often related to secondary deformation, even though that of the matrix seems to be connected with the primary deformation. In these rocks affected by superimposed deformations, finite strain clearly is related to this primary deformation. Thus, the prospect of finite strain calibration by magnetic analysis must be based here upon precise determination of matrix anisotropy. The relation [Mr,

1] = [A,1] A? + [B,1] shows that it is sufficient to determine, either the value of a single term M,1 (for example, by analysis in variable field and temperature—Rochette, 1983), or of the A? value (for example, by a physical separation of rock components—Henry and Daly, 1983) to attain such a determination.

Acknowledgements I am very grateful to John Thompson, Lucien Daly and Michel Westphal for their help in this study.

References Borradaile, G.J., 1973. Dalradian structure and stratigraphy of the Northern Loch Awe district, Argyllshire. Trans. R. Soc. Edinburgh, 69: 1—21. Borradaile, G.J., 1979. Strain study of the Caledonides in the Islay region, SW Scotland. J. Geol. Soc. London, 136: 77-88. Coward, M.P. and Whalley, J.S., 1979. Texture and fabric studies the Kishorn near Kyle of Lochalsn, Westernacross Scotland. J. Struct.Nappe, Geol., 1: 259—273. Daly, L., 1967. Possibilité d’existence dans les roches de plusieurs anisotropies magnCtiques superposées; leur separation. C.R. Acad. Sci. Paris, B 264: 1377—1380. Daly, L., 1970. Etude des proprietes magnétiques des roches métamorphiques ou simplement tectonisées. Nature de leur aimantation naturelle. Determination de leur anisotropie magnétique et application a l’analyse structurale. Thesis, Paris, 340 pp. Daly, L. and Henry, B., 1983. Separation d’anisotropies magnétiques composantes en vue d’applications a l’étude quantitative de Ia deformation des roches. C.R. Acad. Sci. Paris, II 296: 153—156. Downie, C., Lister, T.R., Harris, A.L. and Fettes, D.J., 1971. A palynological investigation of the Dalradian rocks of Scotland. Inst. Geol. Sci., Rep. 71/9, 30 pp. Flinn, D., 1962. On folding during three dimensional progressive deformation. Q. J. Geol. Soc. London, 118: 385—433. Graham, J.W., 1954. Magnetic susceptibility anisotropy: an unexploited element of petrofabric. Geol. Soc. Am. Bull., 65: 1257-1258. Henry, B., 1973. Studies of microtectonics, anisotropy of magnetic susceptibility and paleomagnetism of the Permian Dome de Barrot (France): paleotectonic 17: and61—72. paleosedimentological implications. Tectonophysics, Henry, B., 1974. Sur l’anisotropie de susceptibilité magnétique du granite recent de Novate (Italie du Nord). C.R. Acad. Sci. Paris, D 278: 1171—1174. Henry, B., 1980. Contribution a l’étude des propnCtés magnétiques de roches magmatiques des Alpes. Consequences structurales, regionales et gCnérales. Thesis, Paris, 362 pp. Henry, B., 1983. Interpretation quantitative de l’anisotropie de susceptibilité magnétique. Tectonophysics, 91: 165—177. Henry, B. and Daly, L., 1983. From qualitative to quantitative magnetic anisotropy analysis: the prospect of finite strain calibration. Tectonophysics, 98: 327—336. Hrouda, F., 1979. The strain interpretation of magnetic amsotropy in rocks of the Nisky Jesenik Mountains (Czechoslovakia). Sbor. Geol. ved, UG, 16: 27—62. Hrouda, F., 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophys. Surv., 5: 37—82. Hrouda, F., Stephenson, A. and Woltar, L., 1983. On the standardization of measurements of the anisotropy of magnetic susceptibility. Phys. Earth Planet. Inter., 32: 203—208. Kligfield, R., Owens, W.H. and Lowrie, W., 1981. Magnetic susceptibility anisotropy, strain, and progressive deformation in Permian sediments from the Maritime Alps (France). Earth Planet. Sci. Lett., 55: 181—189.

200 Kligfield, R., Lowrie, W. and Pfiffner, O.A., 1982. Magnetic properties of deformed oolitic limestones from the Swiss Alps: the correlation of magnetic anisotropy and strain. Eclogae Geol. HeIv., 75: 127—157. Kneen, S.J., 1976. The relationship between the magnetic and strain fabrics of some haematite-bearing welsh slates. Earth Planet. Sci. Lett., 31: 413—416. Rathore, J.S., 1979. Magnetic susceptibility anisotropy in the Cambrian slate belt of North Wales and correlation with strain. Tectonophysics, 53: 83—97. Rathore, J.S., 1980. The magnetic fabric of some slates from the Borrowdale volcanic group in the English Lake District and their correlation with strain. Tectonophysics, 67: 207—220.

Rathore, J.S. and Henry, B., 1982. Comparison of strain and magnetic fabrics in Dalradian rocks from the southwest Highlands of Scotland. J. Struct. Geol., 4: 373—384. Roberts, J.L., 1974. The structure of the Dalradian rocks in the southwest Highlands of Scotland. J. Geol. Soc. London, 130: 93—124. Rochette, P., 1983. Propriétes magnétiques et deformations dans des roches sédimentaires alpines. Application au Dogger de la zone dauphinoise. Thesis of speciality, Grenoble, 136 pp. Wood, D.S., Oertel, G., Singh, J. and Bennet, H.P., 1976. Strain and anisotropy in rocks. Philos. Trans. R. Soc. London, Ser. A: 283: 27—42.