Analysis of Variscan dynamics; early bending of theCantabria^Asturias Arc, northern Spain

J.M. Kollmeier, B.A. van der Pluijm, R. Van der Voo *Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109, USA

Received 28 January 2000; accepted 6 June 2000

Abstract

Calcite twinning analysis in the Cantabria^Asturias Arc (CAA) of northern Spain provides a basis for evaluatingconditions of Variscan stress and constrains the arc's structural evolution. Twinning typically occurs during earliestlayer-parallel shortening, offering the ability to define early conditions of regional stress. Results from the Somiedo^Correcilla region are of two kinds: early maximum compressive stress oriented layer-parallel and at high angles tobedding strike (D1c1) and later twin producing compression oriented sub-parallel to strike (D2c1). When all D1compressions are rotated into a uniform east^west reference orientation, a quite linear, north^south trending fold^thrust belt results showing a slight deflection of the southern zone to the south^southeast. North^south-directed D2c1

compression was recorded prior to bending of the belt. Calcite twinning data elucidate earliest structural conditions thatcould not be obtained by other means, whereas the kinematics of arc tightening during D2 is constrained bypaleomagnetism. A large and perhaps protracted D2c1 is suggested by our results, as manifested by approximately 50%arc tightening prior to acquisition of paleomagnetic remagnetizations throughout the CAA. Early east^westcompression (D1c1) likely resulted from the Ebro^Aquitaine massif docking to Laurussia whereas the north-directedcollision of Africa (D2c1) produced clockwise bending in the northern zone, radial folding in the hinge, and rotation ofthrusts in the southern zone. ß 2000 Elsevier Science B.V. All rights reserved.

Keywords: Hercynian orogeny; calcite; twinning; oroclines; stress

1. Introduction

The highly curved Cantabria^Asturias Arc(CAA) of northern Spain forms the core of thegreater Ibero^Armorican Arc as it traces westernVariscan zonations through France (Brittany)into northern Spain (Fig. 1). The Somiedo struc-

tural unit, also known as the fold nappe belt [1],marks the ¢rst occurrence of unmetamorphosedPaleozoic sedimentary units east of the basementcored Narcea Antiform [2]. Deformation contin-ued eastward (toward the foreland) producing theSobia unit, the Central Coal Basin (C.C.B.) [3],Ponga Unit [4], and the more easterly Picos deEuropa (P.D.E.) and Pisuerga^Carrion province(P.C.P). These individual structural units vary incharacter commensurate with variations in lithol-ogy and changing deformation conditions. Basedon deformation style and structural orientation,

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 2 0 3 - X

* Corresponding author. Tel. : +1-734-764-8322;Fax: +1-734-763-4690; E-mail: voo@umich.edu

EPSL 5559 2-8-00

Earth and Planetary Science Letters 181 (2000) 203^216

www.elsevier.com/locate/epsl

we have divided the sampled Somiedo into threezones: a northern zone (NZ), hinge zone (HZ),and southern zone (SZ).

The kinematics of arc formation, particularlyarcuate foreland fold^thrust belts, has gained con-siderable interest since the advent of the platetectonics [5^11]. Unraveling arc development re-quires contributions from as many analytical ap-proaches as possible. However, despite extensivework in the CAA (structural, stratigraphic, pale-omagnetic, seismic, and paleotemperature analy-ses), critical questions persist concerning the ear-liest conditions of foreland deformation, whichcorrespond on a larger scale to the overall char-acter and coeval tectonism in the convergencezone between Laurussia and Gondwana, whichproduced the Ibero^Armorican Arc.

Tectonostratigraphic evidence indicates that in-itial deformation was foreland propagating, be-ginning with the rear-most Somiedo structuralunit in Westphalian B times and a¡ecting theeastern-most structural units in the upper Stepha-nian ([12], and references therein). Spatially, awest to east geometric gradation can be seen

with arcuate structural trends, averaging to anorth^south trend, that gives way to ultimatelyeast^west trending structures in the P.D.E. unit.This change in arc character has been interpretedwith two end-member models: (1) an originallylinear fold^thrust belt that was later bent intoits present geometry (e.g. [13]), and (2) a fold^thrust belt that evolved into today's structureunder a continuously changing stress ¢eld (e.g.[12]). Other interpretations have suggested an ini-tial arc that is partially curved and later tightened[14^16] based on unfolding the radial fold set (re-sponsible for an aerial reduction of 35^40%), dis-placement directions, and agreement with earlier,somewhat erroneous paleomagnetic conclusions.Paleomagnetic studies have generally concludedthat (at least) half of the present curvature is sec-ondary [17^21], rendering this an `orocline' as de-¢ned by Carey [22]. Pares et al. [20] reevaluatedpublished Paleozoic paleomagnetic data by apply-ing inclination-only tilt tests and concluded thatall remagnetizations are syn-tilting to post-tiltingin age, thereby changing the earlier paleomagneticconclusions to some degree. Results from Devon-

Fig. 1. Structural sketch of the Cantabrian Arc showing the traces of principal thrusts and the traces of axial fold planes that re-sulted from longitudinal F1 (closed) and radial F2 (open) fold sets. The Somiedo structural unit has been divided into threezones (NZ, HZ, and SZ), based on variation in structural style and orientation.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216204

ian limestones speci¢cally, reveal that magnetiza-tions were recorded during and after early tiltingand were a¡ected by late vertical-axis rotations[20,23]. Thus the more recent paleomagnetic ¢nd-ings indicate that the horseshoe-shaped CAA re-sulted from late tightening of a partially curved(V50%) fold^thrust belt, leaving the amount ofstructural rotations prior to these remagnetiza-tions unknown.

Calcite twinning analysis can o¡er an e¡ectivemethod for quantifying rotations in curved oro-genic belts (e.g. [7,8]) complementary to paleo-magnetic analysis. Previous ¢eld applications ofcalcite twinning analyses in fold^thrust belts in-clude the Alps [7,8,24,25], the Idaho^Wyomingsalient [26] and the Appalachians [27^30]. Exper-imentally deformed limestones show that individ-ual twins record the orientation of the local stateof stress as well as the orientation and magnitudeof the induced intragranular strain. It is consid-ered that stress recorded over the whole aggregateof a sample is representative of the regional stressstate. Calcite twinning requires a low critical re-solved shear stress (V10 MPa) [31,32] and is astrain hardening process where further twinning isresisted as beds tilt during folding [29,30,33,34].Therefore, it typically preserves early horizontalcompression prior to folding as layer-parallelshortening [27,35]. In these instances, a regionalpaleostress direction can be statistically deter-mined for a sample from the array of paleostressdirections of an aggregate of twinned calcitegrains.

The primary objective of this paper is to deter-mine the original geometry of the CAA that bestsatis¢es paleostress data from deformed lime-stones, in addition to other constraints (paleo-magnetic and regional structural data). Themicrostructural analysis presented here independ-ently evaluates the kinematics of whole arcformation and provides information on thestate of regional stress during Variscan deforma-tion.

2. Geologic background and sampling

The CAA forms the eastern margin of the west-

ern Variscan orogen and is characterized by fore-land-directed concavity. Two fold sets have beenidenti¢ed: (1) early longitudinal folds (tangentialto the arc) overprinted by (2) a cross-cutting ra-dial fold set, forming regional (Type I) interfer-ence structures [36]. Tracing the longitudinalstructure of the Somiedo unit from north tosouth, ¢rst-order folds give way to an increasingnumber of thrusts (Fig. 1). The SZ is not onlyunique in structural style, but also shows amarked change in orientation to an overall east^west trend (with respect to the HZ; Fig. 1).

Paleozoic deposition through Carboniferoustimes resulted in a stable shelf sequence, generallythicker in the west (present-day coordinates) andtapering to the east. Although minor detachmentzones occur in several structurally weak unitsthroughout the lower Paleozoic sequence, the pri-mary basal detachment soles below the CambrianLancara Formation [1]. Two well-exposed Devon-ian limestone units, the Santa Lucia and Portillaformations, were sampled to determine paleo-stress/strain as recorded by the mechanical twin-ning of calcite. These units represent conditions ofa reef-rimmed carbonate platform allowing forthe deposition of bioclastic grainstones and pack-stones with limey mudstones present [37].

Paleotemperature estimates based on conodontalteration index (CAI) values indicate maximumtemperatures of 150^250³C in the Somiedo unit[38,39]. Values associated speci¢cally to the unitssampled in this paper, the Santa Lucia and Por-tilla formations, revealed temperatures in therange of 120^155³C and 70^95³C respectively[38]. This agrees with the observed twinning char-acter of our samples (Fig. 2), which suggests tem-peratures of 9 200³C [40^42].

A concern in deciphering regional paleostressconditions is to avoid e¡ects that may be associ-ated with localized deformation. Thus, weavoided sampling near major faults, areas ofhigh shear and fold hinges, following the recom-mendations of Friedman and Stearns [43]. In par-ticular we sampled the thickest available coarse-grained beds using a portable gasoline-poweredcoring drill. Rock cores were oriented using aBrunton compass mounted on an aluminumcore orienting device.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 205

3. Calcite twinning analysis

3.1. Data collection

Optical measurements de¢ne the crystallo-graphic orientation of the twinned calcite grain(Fig. 3) as well as the twin set(s) measured. Twin-ning occurs most readily on e-twin planes in anyof three orientations (twin sets) per given calcitecrystal. All samples that were analyzed containedcalcite grains that were not preferentially orien-tated crystallographically, thus ensuring that theresults are not a consequence of an intrinsic fabricin the sample. Measured samples range from bio-clastic grainstone to packstone containing crin-oids, corals, bryozoa, and bioturbation observedin an otherwise crystalline calcite matrix. Calcitegrains were observed with one, two or all threetwin sets present, but most commonly contain twosets. For each grain, only twin sets that arestraight and continuous within grains were mea-sured to obtain the most accurate results. Whennecessary, this resulted in measurement of onlyone (acceptable) twin set per given calcite grain.

3.2. Analytical techniques

The use of twinned calcite as a paleostress/strain indicator in both experimentally and natu-rally deformed limestones has been shown to yieldrobust data [7,8,29,34,43^45]. Several methods areavailable to quantify paleostress and paleostrainwith calcite twinning analysis [41,46]. To obtaincompression and tension axes for each sample, weuse the dynamic analysis of Turner [47]. For eachtwinned calcite grain, compression and tensionaxes (C and T, Fig. 3a) are situated 71.5 and18.5³ to the c-axis respectively, and at 45³ to thee-twin plane (Fig. 3a,b). Calcite crystals havethree potential e-twin planes (Fig. 3c), where twin-ning glide will occur preferentially along oneplane depending on the orientation of the remotestress. Initial data were cleaned (removing all re-sidual values (RV), see below) as suggested byGroshong et al. [48], such that populations of 25or more twin sets remain. These are statisticallysigni¢cant to characterize the true twinning strainfor a whole sample [48]. We have chosen the samenumber of twin sets as a minimum number for

Fig. 2. A representative photomicrograph showing twinned calcite grains characterized by multiple (usually two) twin sets andthin twins indicating low temperature conditions.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216206

computing principal stress axes as well. The re-sulting compression and tension axes for a sampleare calculated as point means assuming a Fish-erian distribution. The data are listed in Table 1.

We have applied Groshong's [49] technique todetermine distributions of strain within each sam-ple using the calcite strain gauge program ofEvans and Groshong [50]. Shear strain is deter-mined for individually measured twin sets, fromwhich a complete strain tensor for the whole sam-ple is statistically calculated and assigned by aleast squares solution (multiple linear regression)(see [49] for a full description). If limestone ishomogeneously deformed, twin sets should yielda single (positive) shear strain, which is termed the`expected value' (equivalent to PEV in [49]). How-ever, when a signi¢cant number of twinned grainsis unfavorably oriented, the sample may have ex-perienced multiple homogeneous or inhomogene-ous strain. This results in some proportion of twinsets with opposing shear sense, referred to here as`residual values' (RV). Thus, each grain records

an expected value (EV) (shear strain) if favorablyoriented, or an RV if not [49,50]. Assigning EVsand RVs depend on a twin set's orientation rela-tive to the statistically determined direction ofshear assigned for the bulk of the aggregate [49].

Experimental work of Friedman and Stearns[43] and Teufel [34] have shown that two discretedeformations may be identi¢ed using the dynamicanalysis technique of Turner [47] and the straingauge technique of Groshong [49]. Friedman andStearns [43] concluded that superposed states ofstress are distinguishable if compression axes ofthe dominant twin sets (EVs) are plotted sepa-rately from subordinate twin sets (RVs). Experi-ments by Teufel [34] indicate that larger errors areobserved both in strain magnitude and orienta-tion for samples recording two deformations.Also, a second deformation is most accuratelydetermined when oriented at a high angle to the¢rst ; however obliquely superposed deformationscan be clearly determined as well.

In presenting our data, a number of qualityconsiderations have been included in Table 1.The ¢rst indicator of reliability is a low percent-age of RV from the calcite strain gauge analysis,which re£ects a largely homogeneous positiveshear strain recorded by the sample. When a sam-ple contains a high percentage of RVs, it is splitinto two distinct populations of twin sets (onecomposed entirely of EVs and the other ofRVs), and each are analyzed as separate deforma-tions. If the separation yields low percentages(6 15% is used as a threshold following [34]) ofnewly calculated RVs, then both deformations arelikely to be robust. Additionally, because the ex-pected shear strain for the aggregate is calculatedby a multiple linear regression of the data, thestandard error is an additional measure of dataquality [34]. Lastly, because whole-sample princi-pal stress axes are calculated using a Fisherianstatistical point mean of individual twin set stressaxes, an K95 value is included (Table 1).

4. Results

Of the 33 sites measured from along the Somie-do unit, 26 record D1c1 paleostress orientations.

Fig. 3. (a) A calcite grain with a single e-twin and the com-pressive and tensile stress (C and T) oriented most favorablyto produce twinning (oriented 45³ to the e-twin plane). Thegeometric relation of e-twins to c-axis is ¢xed. (b) The samesituation as in a but now illustrated with an equal area pro-jection (lower hemisphere) to show the orientations of stressas related to the crystallography. (c) All three possible e-twinplanes and their poles are represented.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 207

Tab

le1

Cal

cite

twin

ning

resu

lts

for

all

mea

sure

dsa

mpl

es

Sam

ple

Raw

data

Cle

aned

data

Stra

inor

ient

atio

ns(b

rg.,p

lng.

)/E

long

atio

nSt

ress

Ori

enta

tion

s(b

rg.,p

lng.

)B

eddi

ng

N/R

S.E

.M.

N/R

S.E

.M.

e 1e 2

e 3c

3c

1K

95St

rike

and

dip

S3

S0

c1

c3c

0

(ori

gina

l)(t

reat

ed)

(T)

(C)

O:o

vert

urne

dD

1;D

2A

TC

D1

;D2

P3E

49/1

0.37

248

/00.

388

(020

,06)

/2.1

7(1

15,4

2)/3

0.50

(284

,47)

/31.

6731

3,52

204,

159.

618

0,75

w9

202,

318

26P

3R26

/10.

585

25/0

0.50

0(2

68,1

9)/1

.76

(167

,30)

/30.

05(0

26,5

3)/3

1.71

203,

1131

6,65

15.7

180,

75w

2528

8,3

0344

P11

E27

/03.

015

27/0

3.01

5(2

79,0

7)/3

.26

(184

,39)

/0.8

6(0

17,5

0)/3

4.12

235,

4023

,45

18.6

041,

73e

083,

25P

11R

31/4

0.82

827

/20.

759

(351

,56)

/4.0

4(2

28,2

0)/3

1.57

(128

,26)

/32.

4704

0,13

142,

4121

.304

1,73

e14

0,3

31P

1336

/50.

572

31/0

0.32

4(3

27,7

7)/2

.56

(210

,07)

/30.

42(1

18,1

1)/3

2.14

323,

7821

6,04

18.5

219,

65w

4822

2,05

46P

1640

/10.

804

39/0

0.70

7(0

62,0

4)/2

.61

(323

,63)

/30.

24(1

55,2

6)/3

2.37

239,

0414

7,16

9.8

063,

24s

392

148,

308

396

P17

40/4

2.14

936

/22.

190

(208

,11)

/5.8

3(1

12,2

9)/3

0.67

(316

,58)

/35.

1606

6,11

324,

4611

.908

6,70

n-O

369

154,

183

90P

23E

46/2

0.72

044

/40.

644

(226

,70

)/3.

04(3

27,0

4)/3

1.25

(058

,20)

/31.

7817

1,75

321,

1212

.513

4,84

s-O

3337

121,

0533

55P

23R

30/1

0.80

529

/30.

796

(269

,00)

/2.4

2(3

59,0

1)/0

.64

(151

,89)

/33.

0731

2,01

213,

8515

.313

4,84

s-O

321

221,

311

323

P45

335

/00.

602

35/0

0.60

2(2

57,0

3)/2

.31

(349

,32)

/1.1

9(1

63,5

8)/3

3.50

026,

2216

5,61

10.1

109,

55s

346

183,

103

61S0

240

/30.

731

37/0

0.63

7(1

20,1

9)/3

.05

(000

,55)

/30.

03(2

21,2

8)/3

3.02

084,

4024

1,48

10.9

077,

89n-

O3

7812

7,3

123

117

S07

36/1

0.61

435

/00.

627

(058

,08)

/3.0

2(3

25,2

1)/3

0.36

(169

,68)

/32.

6524

6,02

151,

5312

.307

6,46

s3

7915

7,08

387

S09

37/1

1.20

136

/01.

220

(150

,60)

/4.0

5(0

46,0

8)/3

1.05

(312

,29)

/33.

0011

1,58

302,

3213

187,

56w

3230

0,3

2356

S10

34/4

1.39

030

/01.

054

(276

,03)

/1.9

6(1

83,4

8)/0

.76

(009

,41)

/32.

7227

5,01

004,

6111

096,

85n-

O3

5918

3,24

361

S11

38/3

0.43

735

/00.

431

(056

,16)

/2.3

8(3

19,2

2)/0

.25

(179

,62)

/32.

6408

3,20

345,

2215

.110

2,47

n-O

353

219,

213

25S1

252

/10

0.45

042

/00.

311

(126

,46)

/1.9

9(2

87,4

2)/0

.38

(025

,10)

/32.

3713

6,62

044,

0611

.510

4,68

s07

9,57

S20E

36/0

0.74

736

/00.

747

(314

,08)

/3.8

2(2

22,1

8)/3

0.09

(067

,70)

/33.

7332

7,38

137,

5111

.631

1,72

n-O

360,

17S2

0R31

/50.

740

26/3

0.63

7(0

58,1

3)/2

.23

(317

,41)

/30.

54(1

62,4

6)/3

1.69

122,

4331

7,46

1831

1,72

n-O

085,

09S2

1457

/14

0.78

243

/10.

840

(223

,23)

/4.7

7(0

17,6

5)/3

1.07

(129

,09)

/33.

7022

5,26

020,

6311

.415

0,75

n3

522

3,06

321

S221

41/3

0.79

638

/10.

770

(351

,12)

/3.4

2(0

88,1

9)/3

1.04

(242

,56)

/32.

3834

3,07

076,

1611

123,

49n

332

258,

2114

S223

E46

/00.

624

46/0

0.62

4(0

03,3

2)/2

.05

(197

,57)

/0.6

7(0

97,0

6)/3

2.72

109,

5800

8,06

8.9

196,

80e

2520

1,09

25S2

23R

26/0

0.96

626

/00.

966

(091

,13)

/3.1

3(2

05,6

1)/0

.62

(354

,25)

/33.

7519

2,04

095,

5915

.619

6,80

e41

280,

2136

S244

46/7

0.67

139

/00.

656

(071

,31)

/2.7

1(3

38,0

5)/0

.77

(240

,59)

/33.

4709

2,24

225,

5810

.890

,75

s3

6515

2,3

093

92S2

7735

/61.

271

29/1

1.04

4(0

84,2

5)/4

.27

(323

,48)

/31.

34(1

90,3

1)/3

2.94

098,

3834

9,23

13.7

098,

49n

357

169,

233

75S3

4E56

/10.

659

55/1

0.64

4(2

30,3

7)/4

.25

(352

,35)

/30.

87(1

10,3

3)/3

3.38

202,

5204

6,35

9.5

206,

66e

3524

5,01

69S3

4R28

/20.

890

26/2

0.77

6(0

76,4

7)/4

.2(3

22,2

1)/1

.03

(216

,35)

/35.

2305

8,28

213,

6019

.520

6,66

e51

329,

324

85S3

6E47

/30.

920

44/0

0.86

8(2

32,5

0)/3

.64

(055

,40)

/0.7

7(3

24,0

1)/3

4.41

060,

5316

3,10

12.2

169,

47e-

O3

218

1,03

5S3

6R25

/00.

845

25/0

0.84

5(1

49,0

1)/4

.45

(058

,15)

/0.1

4(1

65,7

3)/3

4.58

178,

1807

2,44

18.7

169,

47e-

O14

264,

0320

S43E

34/0

0.91

234

/00.

912

(336

,53)

/2.8

1(0

82,1

1)/3

0.56

(180

,35)

/32.

2534

5,42

207,

3817

.423

0,74

s-O

5927

3,07

97S4

3R29

/11.

216

28/0

1.18

4(2

02,3

9)/2

.97

(094

,21)

/0.6

7(3

43,4

4)/3

3.64

216,

1832

9,51

18.3

230,

74s-

O33

0,3

54S4

435

/70.

710

28/1

0.58

0(1

52,0

1)/4

.13

(059

,83)

/0.4

2(2

42,0

6)/3

4.54

018,

0010

7,72

14.8

185,

87e-

O30

284,

1540

S45

42/5

0.79

537

/10.

715

(238

,31)

/3.2

8(3

57,3

7)/0

.04

(122

,37)

/33.

3124

2,33

080,

5612

.226

0,79

n-O

105

315,

0971

S48E

37/0

0.71

637

/00.

716

(026

,07)

/3.3

1(1

33,6

5)/0

.04

(293

,24)

/33.

3419

3,00

282,

6112

.222

3,88

w68

297,

323

53S4

8R35

/40.

929

31/0

0.87

1(2

86,2

0)/4

.03

(093

,70)

/30.

57(1

95,0

4)/3

3.46

255,

5405

1,32

21.5

223,

88w

5219

1,08

15S4

9E51

/00.

768

51/0

0.76

8(1

85,7

6)/2

.37

(304

,07)

/30.

16(0

35,1

2)/3

2.21

274,

5102

5,14

8.6

227,

48n

5620

2,06

26S4

9R28

/11.

394

27/0

1.29

4(0

09,0

1)/4

.43

(279

,09)

/30.

62(1

07,8

1)/3

3.85

038,

0929

7,55

14.1

227,

48n

7230

5,09

61S6

8E34

/00.

537

34/0

0.53

7(1

13,3

9)/2

.34

(311

,50)

/31.

01(2

11,0

8)/3

1.33

104,

3231

4,53

11.6

308,

75n

001,

09

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216208

Of these 26 samples, 13 record a second twin-pro-ducing deformation D2c1 (Table 1; Fig. 4). Twoadditional samples (P13A, S79A) record only di-rections that agree best with D2c1 based on theorientation of the recorded stress relative to struc-tural trend. The remaining ¢ve samples are notinterpretable for one of two reasons. (1) Compres-sions are oriented at steep angles to bedding(v 25³), indicating twins that probably re£ect lo-cal perturbations or are syn- to post-folding (e.g.sites S12, S43R, Fig. 4). (2) Multiple deformationsmay be present but are not distinguishable in asample, perhaps due to local rotation and tiltingof folds during arc tightening (in that case com-pressions are intermediately oriented betweenstrike normal and strike parallel, e.g. sites P11,S20, S68, Fig. 4).

Samples recording two deformations typicallyindicate D1c1 and D2c1 to be oriented at highangles to each other. It has previously been estab-lished that longitudinal structures (tangential tothe arc) belong to an early deformation that hasbeen overprinted by a later, radial fold-producingevent [1]. This forms the basis by which we inter-pret D1c1 and D2c1 in samples recording twotwinning events. Maximum compressive stress isconsidered D1c1 when oriented (nearly) layer-parallel and at a high angle to strike of foldedbeds (does not fall under uninterpretable category1 above). This occurrence illustrates the character-istic behavior of twinned calcite to record layerparallel paleostresses and strains that are passivelyrotated with subsequent folding within the sameoverall stress ¢eld (compression directed at a highangle to the hinge line). This allows for distinctionof D1c1 compression oriented at a high angle tostrike and D2c1 sub-parallel to strike (does notfall under uninterpretable category 2 above). Oneexception, sample S43, illustrates that D1c1

(S43R) is uninterpretable (category 1, above)whereas D2c1 (S43E) is robust (Fig. 4)

Whether a sample records one or two deforma-tion events appears to be related to its location inthe region (Fig. 5). Samples that record twoevents are nearly ubiquitous in the HZ whereasa single deformation dominates samples of theSZ. In the NZ, ¢ve of seven samples recordD1c1 directions and four preserve D2c1 compres-T

able

1(c

onti

nued

)

Sam

ple

Raw

data

Cle

aned

data

Stra

inor

ient

atio

ns(b

rg.,p

lng.

)/E

long

atio

nSt

ress

Ori

enta

tion

s(b

rg.,p

lng.

)B

eddi

ng

N/R

S.E

.M.

N/R

S.E

.M.

e 1e 2

e 3c

3c

1K

95St

rike

and

dip

S3

S0

c1

c3c

0

(ori

gina

l)(t

reat

ed)

(T)

(C)

O:o

vert

urne

dD

1;D

2A

TC

D1

;D2

S68R

31/4

0.78

327

/00.

716

(336

,41

)/2.

46(1

57,4

8)/3

0.32

(067

,00)

/32.

1532

3,07

101,

3516

.530

8,75

n26

7,13

S69E

40/0

0.91

940

/00.

919

(015

,32)

/3.1

7(1

42,4

5)/0

.15

(266

,29)

/33.

3304

2,60

268,

2210

.208

9,70

n33

8210

9,08

3367

S69R

27/0

1.28

727

/01.

287

(233

,37)

/3.0

1(1

34,1

1)/0

.41

(030

,51)

/33.

4228

2,26

054,

5515

.808

9,70

n3

6620

7,02

337

S71

30/4

1.08

926

/10.

855

(043

,30)

/2.7

6(1

40,1

4)/0

.20

(251

,56)

/32.

9634

7,01

256,

5913

.307

8,64

n-O

377

202,

323

342

S72

30/4

0.93

226

/00.

879

(072

,15)

/3.7

2(3

32,3

1)/3

0.02

(184

,55)

/33.

7007

1,16

295,

7012

.509

7,73

n-O

358

206,

310

338

S75E

49/1

0.40

648

/00.

377

(142

,01)

/2.5

4(2

39,8

2)/1

.08

(052

,08)

/33.

6228

8,62

044,

129.

807

2,74

s33

9906

7,30

3310

9S7

5R27

/20.

939

25/1

0.81

9(0

54,0

9)/4

.13

(160

,59)

/2.4

1(3

20,3

0)/3

6.54

069,

1818

1,70

15.9

252,

74s

9733

6,03

92S7

7E45

/10.

588

44/0

0.54

5(1

59,7

2)/2

.34

(338

,17)

/30.

49(2

48,0

0)/3

1.86

124,

7829

8,13

9.3

140,

84s

3331

124,

2033

52S7

7R31

/40.

855

27/2

0.86

2(2

68,0

7)/2

.94

(360

,21)

/30.

34(1

59,6

8)/3

2.60

298,

2016

4,58

18.1

140,

84s

315

202,

307

342

S79

40/5

0.94

935

/20.

781

(280

,65)

/3.2

2(0

97,2

5)/3

0.46

(187

,10)

/32.

7730

4,72

137,

1811

.213

3,89

s33

3815

1,04

3325

Nis

the

num

ber

oftw

inse

tsan

alyz

edan

dR

isth

enu

mbe

rof

asso

ciat

edR

Vs.

Raw

data

wer

ecl

eane

dby

rem

ovin

gal

lR

Vs,

assu

gges

ted

byG

rosh

ong

[48]

.B

oth

raw

and

trea

ted

data

are

pres

ente

dw

ith

thei

rre

spec

tive

stan

dard

erro

rs(S

.E.M

.).

Dat

aar

ein

clud

edfo

rbo

thth

est

rain

gaug

ete

chni

que

ofG

rosh

ong

[49]

inco

lum

nse 1

,e 2

and

e 3as

wel

las

the

dyna

mic

anal

-ys

isof

Tur

ner

[47]

inco

lum

nsc

3(T

ensi

on),

c1

(Com

pres

sion

),an

dK

95.

Stri

kean

ddi

pan

dco

mpr

essi

ondi

rect

ion

(c1)

afte

rti

ltco

rrec

tion

(AT

C)

are

give

nw

ith

the

prop

erde

viat

ions

(S3

S0,c3c

0)

plot

ted

inF

ig.

8a,b

(D2c

1de

viat

ions

used

inF

ig.

8bar

ein

bold

face

).

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 209

sion as determined by the relationship betweenstress orientations and strike. The results areshown in map view in Fig. 5.

5. Discussion

5.1. D1 folding and thrusting

All three structural zones of the Somiedo unitare unbent to an earliest geometry by rearrangingD1 stresses to an arbitrarily uniform (present-day)east^west reference orientation. What results is an

essentially linear, generally north^south trendingfold^thrust belt (Fig. 6). However, the SZ featuresa mild foreland-progressive counter-clockwise ro-tation in successively younger thrust sheets, whichculminates in a more southeasterly trend in thefrontal portion. This structural de£ection resultsperhaps from a greater foreland translation viathrusting relative to the HZ and NZ. Averagingthe resultant structural trends from the youngestsampled thrust sheets, we derive an orientation of27³ east of south as a best-¢t D1 structural trendfor all SZ samples.

Furthermore, the reconstruction in Fig. 6 is il-

Fig. 4. Equal-area lower-hemisphere plots of all sample results. Contoured compressive stress axes are shown as well as the prin-cipal strain axes calculated with the strain gauge of Groshong [49]. All data are represented in present-day ¢eld coordinates withbedding included. Tabulated D1c1 and D2c1 compressions illustrate the predominance of D2 associated with EVs and D1 associ-ated with RVs. Stress data are contoured with the ¢xed-circle counting method of Kamb, whereas maximum compressive stressdirections are calculated as point means assuming a Fisherian distribution.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216210

lustrated by a series of three equal-area lower-hemisphere projections of D1 maximum compres-sive stress axes (Fig. 7). Stress directions inpresent-day structures (Fig. 7a) are ¢rst untiltedto paleohorizontal (Fig. 7b), and then rotatedabout a vertical-axis to the original geometry pre-sented in Fig. 6 (Fig. 7c). Even with the scatterdue to assigning all SZ sites an average D1 trendof S 27 E (open symbols, Fig. 7c), the resultingalignment illustrates that uniform as well asnearly bedding-parallel compressions are valid as-

sumptions. In a very generalized fashion, Fig. 7illustrates an overall fold and strike test for theentire study area.

5.2. Arc rotations

When considering whether a curved orogenicbelt is either a primary feature (no rotation), oran `orocline' as put forth by Carey [22], plottingstructural trend versus pre-bending paleomagneticdeclination poses a powerful evaluation of rota-

Fig. 5. Map of the CAA adapted from 1:200 000 scale geological maps [53]. Black arrows indicate the orientation of (tilt-cor-rected) subhorizontal D1c1 and white bars represent later (in situ) D2c1 ; because D2c1 is generally strike parallel, in situ andtilt-corrected orientations are approximately the same. The three structural zones are delineated and labeled as NZ, HZ, and SZ.See Fig. 4 for a more detailed stress to strike relation as local structures may deviate from the generalized trends (e.g. S214).

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 211

tion [9,20,51]. We use an analogous relationshipin Fig. 8, where the axes represent the deviationfrom a reference strike (abscissa) and deviationfrom a reference D1 maximum compressive stress(ordinate). This plot can then be interpreted interms of relative rotations, with the assumptionthat all stress was uniform within a single event.The likelihood of this assumption is later eval-uated by comparing the plots of both deformationepisodes (D1 and D2). Reference values (S0 and

Fig. 6. Arc geometry during D1c1 constrained by untighten-ing the arc to a reference east^west stress ¢eld. North^southtrending thrusts in the SZ have swept eastward in a forelandprogressive sequence as shown by arrows. Thin, short hori-zontal lines (doubly tipped with inward arrows) representcompressive stress and are related with the structural trend(thicker intersecting lines) for all 26 D1 compressions.

Fig. 7. Equal area lower hemisphere projections representingD1c1 orientations of all samples considered to unbend theCAA. Samples from the NZ and HZ are indicated with solidcircles and SZ samples are denoted with open circles. Orien-tations from in situ data (a) are untilted to paleohorizontal(b). Finally, the data are rotated to a reference north trend-ing structure for NZ and HZ samples, and an average struc-tural trend of S27E for all SZ samples (c). The scatter in theSZ data illustrates the variation in stress directions in succes-sive thrust sheets that is comparable to the NZ and HZ scat-ter combined.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216212

c0 in Fig. 8a,b) are chosen merely to produce abest-¢t line that passes through the origin whenconsidering all samples. Plotted points representthe deviation from these reference values (S03S,c03c), and data are presented with maximumcompressive stress axes untilted to paleohorizon-

tal (Table 1). Fig. 8a plots D1 paleostress orien-tations of individual sample sites from around thearc. Ideally, if a curved orogen is a primary fea-ture and a uniform stress ¢eld produced the orig-inal curvature, one would expect uniform stressorientations irrespective of the structural trend.This end-member condition would produce aline along the abscissa with zero slope (no devia-tion in c1 orientations). The other end-member isan initially linear belt, where during bending therecorded stress directions are rotated passivelywith the curved geometry; in this case the ex-pected slope of the best-¢t line is 1. Fig. 8a showsthat the results for the CAA correspond nearlyperfectly to the latter end-member model ; i.e.,curvature from an originally straight belt. Notethat this plot does not assess whether an origi-nally linear structure undergoes continuous or ep-isodic rotation, but merely indicates total rota-tion.

5.3. D2 folding and thrusting

In Fig. 8b the deviation from a reference D2maximum compressive stress is plotted. A slope ofV1 is again observed which indicates that D1 arcgeometry was nearly linear and that D2 was re-corded prior to arc bending. The agreement be-tween the rotations required for both D1 (Fig. 8a)and D2 (Fig. 8b) provides conclusive evidence oftightening with nearly all of the curvature beingsecondary and supports the assumption of uni-form compression for each event.

Fig. 8. (a) D1 compression deviations from the mean refer-ence stress (c0 = 244³) of all 26 samples tilt corrected to bed-ding horizontal plotted against strike deviations from thereference strike (S0 = 155³). (b) Likewise, D2 compression de-viations from the mean reference stress (c0 = 176³) of 13samples plotted against strike deviations from a referencestrike (S0 = 171³). In both cases, a linear regression is drawnwith associated correlation coe¤cients R2 and slope m (indi-cating the CAA was oroclinally bent from a nearly straightoriginal con¢guration, m =V1). (c) Paleomagnetic data [20]for the CAA showing deviations in (mean) declination andstrike for each of the three zones (NZ, HZ, and SZ), illus-trating rotation for approximately half of the CAA curva-ture.6

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 213

5.4. Evolution of the CAA

In samples recording two deformations, partic-ularly those of the NZ and HZ, D2c1 is com-monly oriented sub-perpendicular to D1c1 (Figs.4 and 5). In the D1 fold geometry this results inD2c1 compressive stresses averaging to an essen-tially north^south orientation. Thus, analogous toD1 (Fig. 8a), D2c1 twinning occurs prior to thelater (D2) phase of macroscopic deformation ofthe belt (Fig. 8b). Overall, this suggests that D2produced clockwise bending in the NZ, radialfolding in the HZ, and tightening of thrusts inthe SZ. It is also observed that the SZ showsthe least amount of evidence of D2 as the samplesrecord only D1 maximum compressive stress di-rections in all but two samples.

Fig. 8c is a compilation of paleomagnetic datafor the CAA plotting the averages for each struc-tural zone [20]. The calculated slope is V0.5, in-dicating that the overall arc curvature was about50% at the time the pervasive syn-tectonic remag-netization was acquired. In comparison with thepaleomagnetic results, the calcite twinning datatherefore demonstrate that V50% of the curva-ture was attained prior to the remagnetization,but following D2 calcite twinning.

Fig. 9 presents our interpretation of the evolu-tion of the CAA in sequential block diagrams,beginning with £at lying shelf units recordingpre-folding twinning due to a uniform (present-day) east^west compression (D1c1 ; Fig. 9a). Dur-

ing D1, longitudinal structural trends form in anearly linear belt (Fig. 9b). D2c1 re£ects a changein stress ¢eld from east^west to north^south act-ing upon the previously formed D1 structures(Fig. 9c). Lastly, present-day structural geometryre£ects large scale folding (D2, manifested in ra-dial folds and regional curvature) in response tonorth^south compression (Fig. 9d). A tectonic ex-planation for these events would be an east^westcollision between the Ebro^Aquitaine massif andthe Laurussian margin (D1), followed by a great-er, perhaps protracted north-directed collision ofGondwana with Laurussia [52].

Acknowledgements

We thank J.C. Larrasoan¬a and J. Pares fortheir assistance in the ¢eld and J.H. Harris forhis advice in calcite twinning analysis. We alsothank M.A. Evans and R.H. Groshong for useof the CSG-22 program. Thoughtful reviews byD.A. Ferrill, B. Ward, and M. Burkhard im-proved the manuscript and are greatly appreci-ated. Support for this project was from NationalScience Foundation Grant EAR 9705755. In ad-dition, funding for the ¢eldwork came from aGeological Society of America research grantawarded to J.M. Kollmeier as well as the ScottTurner fund, Dept. of Geological Sciences, theUniversity of Michigan.[AC]

Fig. 9. Block diagrams summarizing the evolution of arc tightening beginning with east^west horizontal compression acting on astable margin in (a) and resulting in an arc geometry (b) after D1 folding and thrusting (as in Fig. 6). A later north^south hori-zontal compression (D2) is shown in (c) and leads to clockwise bending in the NZ, radial folding in the HZ, and tightening ofthrusts in the SZ of present-day arc geometry (d).

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216214

References

[1] M. Julivert, Decollement tectonics in the Hercynian Cor-dillera of northwest Spain, Am. J. Sci. 270 (1971) 1^29.

[2] G. Gutierrez-Alonso, Strain partitioning in the footwall ofthe Somiedo Nappe: Structural Evolution of the Narceatectonic window, NW Spain, J. Struct. Geol. 18 (1996)1217^1229.

[3] J. Aller, J. Gallastegui, Analysis of kilometric scale super-posed folding in the Central Coal Basin (Cantabrian zone,NW Spain), J. Struct. Geol. 17 (1995) 961^969.

[4] J. Alvarez-Marron, Three-dimensional geometry and in-terference of fault-bend folds: Examples from the PongaUnit, Variscan Belt, NW Spain, J. Struct. Geol. 17 (1995)549^560.

[5] A.C. Ries, F.R.S. Shackleton, Patterns of strain variationin arcuate fold belts, Phil. Trans. R. Soc. Lond. A283(1976) 281^287.

[6] S. Marshak, Kinematics of orocline and arc formation inthin-skinned orogens, Tectonics 7 (1988) 73^86.

[7] D.A. Ferrill, R.H. Groshong, Deformation conditions inthe northern Subalpine Chain, France, estimated fromdeformation modes in coarse-grained limestone, J. Struct.Geol. 15 (1993) 995^1006.

[8] D.A. Ferrill, R.H. Groshong, Kinematic model for thecurvatrue of the northern subalpine chain, France,J. Struct. Geol. 15 (1993) 523^541.

[9] R. Van der Voo, J.A. Stamatakos, J.M. Pares, Kinematicconstraints on thrust^belt curvature from syndeforma-tional magnetizations in the Lagos del Valle Syncline inthe Cantabrian Arc, Spain, J. Geophys. Res. 102 (1997)10105^10119.

[10] D. Hindle, M. Burkhard, Strain, displacement and rota-tion associated with the formation of curvature in foldbelts; the example of the Jura arc, J. Struct. Geol. 21(1999) 1089^1101.

[11] J. Macedo, S. Marshak, Controls on the geometry offold^thrust belt salients, Geol. Soc. Am. Bull. 111 (1999)1808^1822.

[12] A. Perez-Estaun, F. Bastida, J.L. Alonso, J. Marqu|nez, J.Aller, J. Alvarez-Marron, A. Marcos, J.A. Pulgar, A thin-skinned tectonics model for an arcuate fold and thrustbelt the Cantabrian Zone (Variscan Ibero^ArmoricanArc), Tectonics 7 (1988) 517^537.

[13] L.U. De Sitter, The Hercynian Cantabrian orogen, Mem.Geopaleontol. Univ. Ferrara 1 (1965) 211^225.

[14] M. Julivert, A. Marcos, Superimposed folding under£exural conditions in the Cantabrian zone (HercynianCordillera, northwest Spain), Am. J. Sci. 273 (1973)353^375.

[15] M. Julivert, M.L. Arboleya, A geometrical and kinemat-ical approach to the nappe structure in an arcuate fold-belt : the Cantabrian nappes (Hercynian chain, NWSpain), J. Struct. Geol. 6 (1984) 499^519.

[16] M. Julivert, M.L. Arboleya, Areal balancing and estimateof areal reduction in a thin-skinned fold-and-thrust belt

(Cantabrian zone, NW Spain): constraints on its emplace-ment mechanism, J. Struct. Geol. 8 (1986) 407^414.

[17] N. Bonhommet, P.R. Cobbold, H. Perroud, A. Richard-son, Paleomagnetism and cross-folding in a key area ofthe Asturian Arc (Spain), J. Geophys. Res. 86 (1981)1873^1887.

[18] H. Perroud, N. Bonhommet, Paleomagnetism of theIbero^Armorican arc and the Hercynian orogeny in west-ern Europe, Nature 292 (1981) 445^447.

[19] A.M. Hirt, W. Lowrie, M. Julivert, M.L. Arboleya, Pale-omagnetic results in support of a model for the origin ofthe Asturian Arc, Tectonophysics 213 (1992) 321^339.

[20] J.M. Pares, R. Van der Voo, J.A. Stamatakos, A. Perez-Estaun, Remagnetization and postfolding oroclinal rota-tions in the Cantabrian/Asturian Arc, northern Spain,Tectonics 13 (1994) 1461^1471.

[21] S.A. Stewart, Paleomagnetic analysis of fold kinematicsand implications for geological models of the Cantabri-an/Asturias Arc, north Spain, J. Geophys. Res. 100 (1995)20079^20094.

[22] S.W. Carey, The orocline concept in geotectonics, Proc.R. Soc. Tasman. 89 (1955) 255^289.

[23] A.B. Weil, R. Van der Voo, B.A. van der Pluijm, J.M.Pares, The formation of an orocline by multiple deforma-tions: A paleomagnetic investigation of the Cantabria^Asturias Arc (Cantabria Zone, northern Spain), J. Struct.Geol. 22 (2000) 735^756.

[24] D. Dietrich, H. Song, Calcite fabrics in a natural shearenvironment, the Helvetic nappes of western Switzerland,J. Struct. Geol. 6 (1984) 19^23.

[25] D. Dietrich, Calcite fabrics around folds as indicators ofdeformational history, J. Struct. Geol. 8 (1986) 655^668.

[26] J.P. Craddock, A. Kopania, D.V. Wiltschko, Interactionbetween the northern Idaho^Wyoming thrust belt andbounding basement blocks, central western Wyoming,Geol. Soc. Am. Mem. 171 (1988) 333^351.

[27] T. Engelder, The nature of deformation within the outerlimits of the central Appalachian foreland fold and thrustbelt in New York State, Tectonophysics 55 (1979) 289^310.

[28] D.V. Wiltschko, D.A. Medwede¡, H.E. Millson, Distri-bution and mechanisms of strain within rocks on thenorthwest ramp of the Pine Mountain block, southernAppalachian foreland: A ¢eld test theory, Geol. Soc.Am. Bull. 96 (1985) 426^435.

[29] B. Kilsdonk, D.V. Wiltschko, Deformation mechanismsin the southeastern ramp region of the Pine Mountainblock, Tennessee, Geol. Soc. Am. Bull. 100 (1988) 653^664.

[30] J.H. Harris, B.A. van der Pluijm, Relative timing of cal-cite twinning strain and fold^thrust belt development;Hudson Valley fold^thrust belt, New York, USA,J. Struct. Geol. 20 (1998) 21^31.

[31] W.R. Jamison, J.H. Spang, Use of calcite twin lamellae toinfer di¡erential stress, Geol. Soc. Am. Bull. 87 (1976)868^872.

[32] H.R. Wenk, T. Takeshita, E. Bechler, B.G. Erskine, S.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216 215

Matthies, Pure shear and simple shear calcite textures.Comparison of experimental, theoretical and naturaldata, J. Struct. Geol. 9 (1987) 731^745.

[33] F.A. Donath, L.S. Fruth, Dependence of strain rate ef-fects on deformation mechanisms and rock type, J. Geol.79 (1971) 347^371.

[34] L.W. Teufel, Stain analysis of experimental superposeddeformation using calcite twin lamellae, Tectonophysics65 (1980) 291^309.

[35] A.A. Chinn, R.H. Konig, Stress inferred from calcite twinlamellae in relation to regional structure of northwest Ar-kansas, Geol. Soc. Am. Bull. 84 (1973) 3731^3736.

[36] J.G. Ramsey, The deformation of early linear structuresin areas of repeated folding, J. Geol. 68 (1960) 75^93.

[37] T.J.A. Reijers, Devonian basin-¢ll histories of the SpanishCantabrian Mountains and the Belgian Ardennes; a com-parison, Geol. Mijnb. 64 (1985) 41^62.

[38] J.G.M. Raven, B.A. van der Pluijm, Metamorphic £uidsand transtension in the Cantabrian Mountains of north-ern Spain; an application of the conodont colour alter-ation index, Geol. Mag. 123 (1986) 673^681.

[39] F. Bastida, C. Brime, S. Garcia-Lopez, G.N. Sarmiento,Tectono-thermal evolution in a region with thin-skinnedtectonics: the western nappes in the Cantabrian Zone(Variscan belt of NW Spain), Int. J. Earth Sci. 88(1999) 38^48.

[40] D.A. Ferrill, Calcite twin widths and intensities as meta-morphic indicators in natural low-temperature deforma-tion of limestone, J. Struct. Geol. 13 (1991) 667^675.

[41] M. Burkhard, Calcite twins, their geometry, appearanceand signi¢cance as stress^strain markers and indicators oftectonic regime: a review, J. Struct. Geol. 15 (1993) 351^368.

[42] D.A. Ferrill, Critical re-evaluation of di¡erential stressestimates from calcite twins in coarse-grained limestone,Tectonophysics 285 (1998) 77^86.

[43] M. Friedman, D.W. Stearns, Relations between stressesinferred from calcite twin lamellae and macrofractures,Teton Anticline, Montana, Geol. Soc. Am. Bull. 82(1971) 3151^3162.

[44] G.L. Twombly, J.H. Spang, Experimental test of the cal-cite strain gage technique at low con¢ning pressure, Tec-tonophysics 89 (1982) T17^T26.

[45] R.H. Groshong, Experimental test of least square straingage calculation using twinned calcite, Geol. Soc. Am.Bull. 85 (1974) 1855^1864.

[46] M. Friedman, G.M. Sowers, Petrofabrics: A critical re-view, Can. J. Earth Sci. 7 (1970) 477^497.

[47] F.J. Turner, Nature and dynamic interpretation of defor-mation lamellae in calcite of three marbles, Am. J. Sci.251 (1953) 276^298.

[48] R.H. Groshong, L.W. Teufel, C. Gasteiger, Precision andaccuracy of the calcite strain^gauge technique, Geol. Soc.Am. Bull. 95 (1984) 357^363.

[49] R.H. Groshong, Strain calculated from twinning in cal-cite, Geol. Soc. Am. Bull. 83 (1972) 2025^2038.

[50] M.A. Evans, R.H. Groshong, A computer program forthe calcite-strain technique, J. Struct. Geol. 16 (1994) 277^282.

[51] S.Y. Schwartz, R. Van der Voo, Paleomagnetic evaluationof the orocline hypothesis in the central and southernAppalachians, Geophys. Res. Lett. 10 (1983) 505^508.

[52] P. Matte, Tectonics and plate tectonics model for theVariscan belt of Europe, Tectonophysics 126 (1986)329^374.

[53] M. Julivert, J. Truyols, J. Garc|a-Alcalde, Mapa Geolo-gico de Espan¬a E 1:200 000 2nd Serie, Hoja No 10,Meires, Instituto Geologico y Minero de Espan¬a, Madrid,1981.

EPSL 5559 2-8-00

J.M. Kollmeier et al. / Earth and Planetary Science Letters 181 (2000) 203^216216