GEOLOGY | Volume 43 | Number 1 | www.gsapubs.org 55

High-resolution X-ray computed microtomography: A holistic approach to metamorphic fabric analysesMohammad Sayab1, Jussi-Petteri Suuronen2, Pentti Hölttä1, Domingo Aerden3, Raimo Lahtinen1, and Aki Petteri Kallonen2

1Geological Survey of Finland, P.O. Box 96, FI-02151 Espoo, Finland2Department of Physics, University of Helsinki, PO Box 64, 00014 Helsinki, Finland3Departamento de Geodinámica and IACT-CSIC, Universidad de Granada, Granada 18002, Spain

ABSTRACTAn intrinsic limitation of studying microstructures in thin section is that their spatial

(three-dimensional, 3-D) distribution, shape, and orientation have to be inferred by com-bining 2-D data from different sections. This procedure always involves some degree of interpretation that in some cases can be ambiguous. Recent advances in high-resolution X-ray computed microtomography have made possible the direct imaging in 3-D of volumes of rock to centimeter scale. This rapidly evolving technology is nondestructive and provides a holistic approach of microstructural analysis that eliminates interpretative procedures associated with 2-D methods. Spatial images can be generated through any part of the rock sample and used as virtual petrographic sections. Our application of this technique to an oriented drill core sample from the classic Orijärvi metamorphic region of southern Fin-land reveals a number of in situ 3-D aspects, including: (1) the spatial distribution and shape of andalusite porphyroblasts, (2) the geometry of a matrix foliation anastomosing around the porphyroblasts, (3) a millimeter-scale compositional layering that controlled the oscillation of porphyroblasts and sulfide mineralization, and (4) distinct inclusion trail patterns char-acterizing porphyroblast core versus rim zones. The combined data indicate that the steeply dipping bedding-subparallel foliation that characterizes the Orijärvi area formed by bulk north-south crustal shortening and associated vertical stretching.

INTRODUCTIONMuch of our knowledge about crust and

mantle dynamics is based on the study of metamorphic minerals and associated micro-structures. Porphyroblastic microstructures in particular represent a unique record of the pres-sure-temperature evolution of a rock linked to its deformation history (e.g., Vernon, 2004). The large majority of this research is based on the study of petrographic thin sections or polished rock surfaces with optical microscope, electron microscope, or microprobe (e.g., Passchier and Trouw, 2005). An important limitation of these tools is their inability to directly visualize micro-structures in three dimensions (3-D). At best, the spatial geometry of rocks can be approximated via the combination of 2-D data from multiple (thin) sections. The lack of full 3-D control commonly introduces ambiguity in microstruc-tural interpretations. For example, sigmoidal inclusion trails have been frequently interpreted in terms of shearing-induced porphyroblast rotation while it was tacitly assumed that the rotation axes must be normal to the stretching lineation (Kriegsman et al., 1989). In a number of cases, however, later work showed that both elements are in fact parallel or oblique, and an alternative origin of the same microstructures via overgrowth of crenulations was concluded (e.g., Sayab, 2005). In complexly deformed rocks, multiple stages of porphyroblast growth are commonly associated with distinctly ori-ented inclusion trail curvature axes. Their dis-tinction and measurement require integrated

study of 6–8 differently oriented thin sections of samples (Bell et al., 1995; Aerden, 2003) and even then involve some degree of interpre-tation and extrapolation of 2-D data between sections (Aerden et al., 2010). Recent techni-cal advancements have added a promising new tool to existing microstructural methods: com-puted microtomography (CT) with high-energy X-rays. The main advantage of the technique is that it allows metamorphic microstructures and minerals to be directly visualized in 3-D at high resolution (e.g., Denison et al., 1997; Hud-dlestone-Holmes and Ketcham, 2010), thereby eliminating the interpretative procedures associ-ated with conventional methods. This technique is nondestructive and provides detailed 3-D spa-tial imagery of the internal architecture of a rock by measuring the attenuation of X-rays as they pass through different mineral phases (Carl-son and Denison, 1992; Ketcham and Carlson, 2001; Ketcham, 2005). In addition to 3-D spatial images, an unlimited number of serial cross sec-tions can be generated as a new kind of virtual petrographic section. In this paper the potential of this method is illustrated as applied to a drill core sample from the Orijärvi region, southern Finland, precisely where Eskola (1915) devel-oped the concept of metamorphic facies. Before being extracted, the sample was oriented in the field in order to match 3-D microstructural data to the tectonic framework and mineralization history of the study area (Skyttä et al., 2006). Through virtual scrolling, either horizontally or vertically, along or across the foliation using

advanced image processing software, the 3-D shape of metamorphic fabrics can be visualized, and thus provides a new holistic approach for detailed microstructural analysis. The technique allows us to sharply delimit rock volumes with variable compositions in the same sample that then can be separated physically and geochemi-cally analyzed. An alternative nondestructive approach to resolve and segment chemical information in 3-D is combining CT with 2-D micro-X-ray fluorescence imaging (Boone et al., 2011). An additional advantage of CT imag-ing is that it allows us to determine the optimal thin section to cut through a rock. We show how the high-resolution X-ray computed microto-mography (HRXCT) is particularly well suited to resolving the spatial distribution of micro-structural controls on sulfide minerals. Such data extrapolated to regional-scale structures are very relevant to the targeting of ore deposits.

SAMPLE DESCRIPTIONSample O1 is a 2.5-cm-diameter, 14-cm-long

andalusite-mica schist, vertically drilled using a hand-held drilling machine, from the Orijärvi area, southwest Finland (Finnish National Grid coordinates: 6686250, 3308859). While still in situ, the drill core was marked with a north-pointing oriented groove on the top surface so that it could be easily reoriented in the X-ray scanner (Fig. 1). Cylindrical drill core is ideal for the HRXCT analysis as it images a circu-lar field of view, where the X-ray source and detector remain stationary. The drill core sam-ple is characterized by a steeply (78°) south dipping, east-west–striking pervasive folia-tion (S1 in Figs. 1A and 2A) that is at ~30° to S0. The foliation is associated with regionally developed upright folds formed as a result of a broadly north-south–directed shortening phase of the early Svecofennian orogeny dated as ca. 1875 Ma (Skyttä et al., 2006). In outcrop, the main S1 foliation can be seen to be overprinted by a widely spaced, subvertical S2 foliation striking northeast-southwest, but this fabric is hardly recognizable in the studied drill core (Fig. 1A). The regional distribution of anda-lusite, cordierite, and fibrolitic sillimanite in the Orijärvi area indicate low-pressure, high-temperature amphibolite facies metamorphic conditions, where andalusite to sillimanite progression is reported (Eskola, 1915; Skyttä et al., 2006).

GEOLOGY, January 2015; v. 43; no. 1; p. 55–58 | doi:10.1130/G36250.1 | Published online 21 November 2014

© 2014 Geological Society of America. For permission to copy, contact editing@geosociety.org.

56 www.gsapubs.org | Volume 43 | Number 1 | GEOLOGY

Two phases of andalusite porphyroblasts have been recognized in the horizontally and verti-cally oriented thin sections (And 1, And 2; Fig. 2B). Porphyroblast rims include well-aligned trails of needle-shaped inclusions that are gen-erally continuous with the intensely developed matrix foliation. Porphyroblast cores are more densely populated with mineral inclusions, but these are mainly equidimensional and do not

exhibit a preferred orientation. Tectonic impli-cations of these textures are discussed herein.

HIGH-RESOLUTION X-RAY COMPUTED MICROTOMOGRAPHY

HRXCT can precisely image the interior of solid materials such as rocks. In contrast to medical X-ray computed tomography, the small X-ray source size and/or smaller detector pixels used in microtomography permit higher reso-lution, and longer exposure times are possible because the irradiated material is inanimate. The scanner generates a series of grayscale radiographs of a given rock sample that are then reconstructed into a 3-D volumetric image of the internal structure of the sample. The gray value of each cubic volume element, or voxel (cf. pixel: picture element), reflects the relative linear X-ray attenuation coefficient, which is dependent on the density and average atomic number of the mineral, and X-ray energy. The 3-D volume can be viewed as individual cross-sectional images (slices) or a 3-D rendering, and quantitatively analyzed with 3-D image process-ing techniques. A more detailed and technical account of CT and its applications to geological materials can be found in Ketcham and Carlson (2001) or Cnudde and Boone (2013).

The HRXCT used in this study is the Nanotom 180 (Phoenix|X-ray Systems and Services, Ger-many; now part of GE Measurement Systems and Solutions) hosted in the University of Hel-sinki Department of Physics (Fig. 1B). X-rays from a tungsten target were used with the X-ray tube voltage set to 160 kV, the beam current set to 120 µA, and 0.5 mm of copper used to filter the X-ray beam. We acquired 1440 views per 360°, with 8 s total exposure time per view. Two segments of the drill core sample (O1-A and O1-B) measuring 2.5 × 5 cm were scanned sep-arately (Fig. 1B). The 3-D images are composed of 2000 horizontal slices from each drill core segment with a voxel size of 14 × 14 × 14 mm. Minerals of interest were manually delineated in approximately every 30th slice, based on the

gray values and texture in the 3-D image, and segmented by interpolating the selections. The resulting images were double-checked using reflected and polarized light microscopes in order to precisely determine minerals (Figs. 2A and 3A). Horizontal elongated holes generated in the matrix are due to anastomosing effects of the main matrix foliation around uneven sur-faces of andalusite porphyroblasts (Fig. 4A).

3-D VISUALIZATION AND ANALYSESThe CT data were processed using Avizo Fire

software (www.fei.com/software/avizo-fire-ms-brochure.pdf) with built-in algorithms for 3-D visualization and rendering. Contrasting gray-scale values allowed us to segment and sepa-rate andalusite porphyroblasts and their inclu-sions, sulfides, quartz, and mica in the matrix. The brightest grains are sulfides, followed by mica, and the darkest are andalusite and quartz (Fig. 3). North-south and east-west vertical sec-tions (Figs. 3B and 3C, respectively) cutting andalusite porphyroblast rims reveal steeply pitching S1 inclusion trails subparallel to matrix S1. Vertical and horizontal slices oriented per-pendicular to the matrix foliation exhibit tight F1 folding of relict sedimentary bedding (S0) with steeply south dipping axial planes parallel to S1 (Fig. 3B) and moderately east plunging fold axes (Fig. 3F).

In all horizontal slices of both the O1-A and O1-B segments of the drill core, andalusite porphyroblasts are preferentially aligned west-northwest–east-southeast (Figs. 3A and 3D), whereas the anastomosing matrix foliation (S1) is deviated around them (Figs. 3E, 4A, and 4B). The horizontal thin section of the sample exhib-its similar textural relationships (Fig. 2B). The 3-D imagery evidences two separate west-north-west–east-southeast–striking layers within the drill core. The southern layer lacks porphyrob-lasts, but contains numerous small sulfide grains (~100–1000 µm) that are preferentially distrib-uted along the matrix foliation (Figs. 4B and 4C). The northern layer hosts andalusite por-phyroblasts, but only scarce sulfide grains (Figs. 3B, 3D, 4B, and 4C). The two layers that prob-ably represent sedimentary bedding were physi-cally separated for detailed chemical analysis by X-ray fluorescence (Table 1); results show the marked differences in Al

2O3, CaO, Na2O, and S contents in the northern and southern layers.

Andalusite porphyroblasts are mostly tabular but show strong elongation in vertical direction (Figs. 4E and 4F). The largest porphyroblast is 18 × 5 × 4 mm and has a core containing unori-ented inclusions surrounded by a rim with well-aligned inclusions (Fig. 4E). This microstruc-ture indicates that prior to development of the steeply dipping foliation during F1 folding, no metamorphic fabric, or a very weak one, existed in the rock. Thus, the andalusite porphyrob-lasts grew mostly prior to development of the

SampleX-ray detector X-ray tubeB

O1-A

O1-B

A

5 cm

S0

S2

S2

N

10 cm

5 mm

NB

And1

2And

S1

S0

S1

1 mm

A

B

Figure 1. A: Field photo (looking down) showing S1 subparallel to S0 cut by spaced S2. B: X-ray computed microtomography setup. Two separate segments (O1-A and O1-B) of the drill core were scanned.

Figure 2. A: Oriented photomicrograph of horizontal thin section through the drill core. The principle matrix foliation (S1) is oblique to west-northwest–east-south-east–striking bedding (S0), which appar-ently controlled the location of andalu-site porphyroblasts. B: Andalusite cores (And-1) preserve unoriented, mostly equi-dimensional, inclusions, whereas rims (And-2) contain well-aligned inclusions that can be followed into the matrix.

GEOLOGY | Volume 43 | Number 1 | www.gsapubs.org 57

regional S1 cleavage. This fabric progressively intensified against porphyroblast margins and eventually included late-stage andalusite rims (Aerden et al., 2010).

DISCUSSION AND CONCLUSIONSUntil 1990, the majority of workers studied

thin sections that were not precisely oriented

relative to geographic coordinates and were cut either perpendicular and or parallel to dominant matrix fabrics. This approach changed when Hayward (1990) introduced a technique for determining the orientation of crenulation axes preserved within porphyroblasts and matrix from radial sets of vertical thin sections of single samples. The method was further refined

by Bell et al. (1995), and its application since then has allowed us to resolve the tectonometa-morphic histories of numerous mountain belts in unprecedented detail. A closely related com-puter technique, FitPitch (developed by Aerden, 2003), allowed calculation of preferred orienta-tions of internal foliations (inclusion trails) from pitch and strike measurements collected in sets of differently oriented thin sections. HRXCT has already been widely used for 3-D visual-ization of igneous (e.g., Jerram et al., 2009), metamorphic (e.g., Huddlestone-Holmes and Ketcham, 2010), and ore minerals (Barnes et al., 2008; Liu et al., 2014). Our HRXCT data illuminate the complete internal architecture

Top view

5 mm

N

5 mm 5 mm

B

C

North

5 mm

North

5 mm

Sulfide grains

And

Sulfide grains

And

Mica

And

And

Sulfide grains

North East

5 mm

Perspective view

North

S1

F1 folds

S1

Sulfide grains

S0

S1

S1

A B C

D E F

North

5 mm

And

And

And

And

5 mm

North

sniarg edfiluS

5 mm

NorthMap view

And

5 mm

And

Perspective view

Trails Trails

Matrix strike

anastomosing foliation

North 5 mm

And

And

East

5 mm

Inclusion trails

Sulfide grains

A

D

CB

E F

Figure 4. Three-dimensional (3-D) recon-struction of high-resolution X-ray computed microtomography data. Data are segmented and colored manually to differentiate objects of interest. A: Segment O1-A of the drill core showing anastomosing foliation around andalusite, whereas in the south the folia-tion is straight. Horizontal elongated holes in the matrix formed due to anastomosing effects around uneven surfaces of andalu-site porphyroblasts. B, C: 3-D rendering of andalusite porphyroblasts (tabular shape), matrix foliation, and distribution of sulfide grains in map (B) and perspective view (C). D: Elongate (vertical orientation) andalusite with steeply pitching inclusions. E: Two an-dalusite porphyroblast with steeply pitching inclusion trails in their rim. F: 3-D distribu-tion of andalusites in the lower drill core segment (O1-B).

Figure 3. Different three-dimensional (3-D) representations of high-resolution X-ray computed microtomography data. The brightest grains are sulfides followed by progressively less bright mica, quartz, and andalusite. A: Horizontal slice showing three andalusite porphyroblasts localized in the northern layer versus sulfide grains in the southern layer. Red lines are manu-ally traced and show S1 foliation. B: Verti-cal slice perpendicular to the matrix foliation showing F1 folds. C: Vertical slice parallel to the matrix foliation and through a porphy-roblast rim with steeply pitching inclusion trails. Dotted red line represents geometry of inclusion trails (manually traced). D, E: 3-D rendered volume of the upper segment (O1-A of the drill core). F: Vertical and hori-zontal slices of the lower segment (O1-B) of the drill core showing east-plunging west-northwest–east-southeast–trending folds. Red, green, and purple arrows indicate x, y, and z coordinates, respectively.

TABLE 1. WHOLE-ROCK GEOCHEMICAL COMPOSITION (X-RAY FLUORESCENCE) OF THE TWO SEPARATE LAYERS

Sample Na2O MgO Al2O3 SiO2 P2O5 K2O CaO TiO2 MnO Fe2O3 S

SM-2014–1A 1.28 2.91 21.60 61.70 0.147 3.170 1.94 0.690 0.043 7.25 0.012SM-2014–1B 2.26 2.77 16.90 61.50 0.165 3.030 3.39 0.670 0.038 7.80 0.426

58 www.gsapubs.org | Volume 43 | Number 1 | GEOLOGY

of metamorphic fabrics and minerals in a drill core sample. Chemically and texturally distinct metamorphic layers enriched in either sulfide minerals or andalusite porphyroblasts have been recognized. The sulfide grains are preferentially oriented along the main matrix foliation, demon-strating sulfide remobilization during formation of the S1 foliation in the Svecofennian orogeny, dated as ca. 1875 Ma (Skyttä et al., 2006). Thus, extensive syntectonic remobilization of volca-nogenic massive sulfide–type sulfide deposits in the Orijärvi area probably took place at the time this fabric formed. Because inclusion trails in andalusite cores are unaligned, no foliation, or only a very weak foliation, existed at the time of nucleation. Porphyroblast growth was sub-sequently coeval with the progressive develop-ment of a steeply dipping, east-west–striking anastomosing foliation (S1) reflecting broadly north-south shortening accompanying the meta-morphism early (ca. 1875 Ma) during the Sve-cofennian orogeny (Fig. 5). Porphyroblast rim growth, including S1, probably occurred early during development of the younger S2 north-east-southwest–striking spaced foliation. The Orijärvi domain of the Uusimaa Belt (southern Finland) is characterized by a steep bedding-subparallel tectonic foliation that formed during a broadly north-south crustal shortening phase of the ca. 1875 Ma Svecofennian event (Skyttä et al., 2006), associated with subvertical stretch-ing (Cagnard et al., 2007) and oroclinal bend-ing (Lahtinen et al., 2014). Thus, the elongated tabular shape of andalusite crystals (Figs. 4 and 5) was probably controlled by vertical stretching during this event.

ACKNOWLEDGMENTSDiscussions with Mikko Nironen and Matti

Pajunen of the Geological Survey of Finland (GTK) are acknowledged. Aerden acknowledges research projects P09-RNM-5388 and CGL2010-21048. We thank Ritva Serimaa, Department of Physics, Helsinki University, for extending the computed microtomog-raphy facilities to the GTK, and Bob Holdsworth and three anonymous reviewers for constructive reviews.

REFERENCES CITEDAerden, D.G.A.M., 2003, Preferred orientation of

planar microstructures determined via statisti-cal best-fit of measured intersection-lines: The ‘FitPitch’ computer program: Journal of Struc-tural Geology, v. 25, p. 923–934, doi:10.1016 /S0191 -8141(02)00119-0.

Aerden, D.G.A.M., Sayab, M., and Bouybaouene, M.L., 2010, Conjugate-shear folding: A model for the relationships between foliations, folds and shear zones: Journal of Structural Geology, v. 32, p. 1030–1045, doi:10.1016/j .jsg .2010 .06.010.

Barnes, S.J., Fiorentini, M.L., Austin, P., Gessner, K., Hough, R.M., and Squelch, A.P., 2008, Three-dimensional morphology of magmatic sulfides sheds light on formation and sulfide melt migra-tion: Geology, v. 36, p. 655–658, doi: 10.1130 /G24779A.1.

Bell, T.H., Forde, A., and Wang, J., 1995, A new indi-cator of movement direction during orogenesis: Measurement technique and application to the

Alps: Terra Nova, v. 7, p. 500–508, doi: 10.1111 /j .1365 -3121.1995.tb00551.x.

Boone, M., Dewanckele, J., Boone, M., Cnudde, V., Silversmit, G., Van Ranst, E., Jacobs, P., Vincze, L., and Van Hoorebeke, L., 2011, Three-dimen-sional phase separation and identification in granite: Geosphere, v. 7, p. 79–86, doi: 10.1130 /GES00562.1.

Cagnard, F., Gapais, D., and Barbey, P., 2007, Col-lision tectonics involving juvenile crust: The example of the southern Finnish Svecofen-nides: Precambrian Research, v. 154, p. 125–141, doi:10.1016/j.precamres.2006.12.011.

Carlson, W.D., and Denison, C., 1992, Mechanisms of porphyroblast crystallization: Results from high resolution computed X-ray tomography: Science, v. 257, p. 1236–1239, doi:10.1126 /science.257.5074.1236.

Cnudde, V., and Boone, M.N., 2013, High resolution X-ray computed tomography in geosciences: A review of the current technology and applica-tions: Earth-Science Reviews, v. 123, p. 1–17, doi: 10.1016 /j .earscirev.2013.04.003.

Denison, C., Carlson, W.D., and Ketcham, R.A., 1997, Three-dimensional quantitative textural analysis of metamorphic rocks using high-resolution computed X-ray tomography; Part 1. Methods and techniques: Journal of Metamor-phic Geology, v. 15, p. 29–44, doi:10.1111 /j .1525 -1314 .1997.00006.x.

Eskola, P., 1915, On the relations between the chemi-cal and mineralogical composition in the meta-morphic rocks of the Orijärvi region: Bulletin de la Commission Géologique de Finlande, v. 44, 145 p.

Hayward, N., 1990, Determination of early fold axis orientation in multiply deformed rocks using porphyroblast inclusion trails: Tectonophysics, v. 179, p. 353–369, doi: 10.1016/0040 -1951 (90) 90301-N.

Huddlestone-Holmes, C.R., and Ketcham, R.A., 2010, An X-ray computed tomography study of inclu-sion trail orientation in multiple porphyroblasts from a single sample: Tectonophysics, v. 480, p. 305–320, doi:10.1016/j.tecto .2009 .10.021.

Jerram, D.A., Mock, A., Davis, G.R., Field, M., and Brown, R.J., 2009, 3D crystal size distri-butions: A case study on quantifying olivine populations in kimberlites: Lithos, v. 112, p. 223–235, doi:10.1016/j.lithos.2009.05.042.

Ketcham, R.A., 2005, Three-dimensional grain fab-ric measurements using high-resolution X-ray

computed tomography: Journal of Structural Geology, v. 27, p. 1217–1228, doi:10.1016/j .jsg .2005.02.006.

Ketcham, R.A., and Carlson, W.D., 2001, Acquisi-tion, optimization and interpretation of X-ray computed tomographic imagery: Applications to the geosciences: Computers & Geosciences, v. 27, p. 381–400, doi:10.1016/S0098-3004 (00) 00116-3.

Kriegsman, L.M., Aerden, D.G.A.M., Bakker, R.J., den Brok, S.W.J., and Schutjens, P.M.T.M., 1989, Variscan tectonometamorphic evolution of the eastern Lys-Caillaouas massif, Central Pyrenees—Evidence for late-orogenic exten-sion prior to peak metamorphism: Geologie en Mijnbouw, v. 68, p. 323–333.

Lahtinen, R., Johnston, S.T., and Nironen, M., 2014, The Bothnian coupled oroclines of the Sveco-fennian orogen: A Palaeoproterozoic terrane wreck: Terra Nova, v. 26, p. 330–335, doi: 10.1111 /ter .12107.

Liu, P.-P., Zhou, M.-F., Chen, W.T., Boone, M., and Cnudde, V., 2014, Using multiphase solid inclu-sions to constrain the origin of the Baima Fe-Ti-(V) oxide deposit, SW China: Journal of Petrol-ogy, v. 55, p. 951–976, doi:10.1093 /petrology /egu012.

Passchier, C.W., and Trouw, R.A.J., 2005, Microtec-tonics (second edition): Berlin, Springer-Verlag, 366 p.

Sayab, M., 2005, Microstructural evidence for N–S shortening in the Mount Isa Inlier (NW Queensland, Australia): The preservation of early W–E-trending foliations in porphyrob-lasts revealed by independent 3D measurement techniques: Journal of Structural Geology, v. 27, p. 1445–1468, doi:10.1016/j.jsg.2005.01.013.

Skyttä, P., Väisänen, M., and Mänttäri, I., 2006, Preservation of Palaeoproterozoic early Sve-cofennian structures in the Orijärvi area, SW Finland—Evidence for polyphase strain parti-tioning: Precambrian Research, v. 150, p. 153–172, doi:10.1016/j.precamres.2006.07.005.

Vernon, R.H., 2004, A practical guide to rock micro-structures: Cambridge, UK, Cambridge Uni-versity Press, 606 p.

Manuscript received 7 September 2014 Revised manuscript received 22 October 2014 Manuscript accepted 29 October 2014

Printed in USA

Bedding

S1N

Non-preferred inclusions

Preferred inclusions

Figure 5. Synthetic structural cross section across the Orijärvi-Kuovila region, Finland, where upright folds are cut by steeply dipping S1 foliation (modified after Skyttä et al., 2006; Cagnard et al., 2007). Randomly oriented inclusions in andalu-site cores versus well-aligned inclusions in rims suggest that prior to metamor-phism, only a weak or no metamorphic fabric had developed in the area.