Mével, C, Gillis, K.M., Allan, J.F., and Meyer, P.S. (Eds.), 1996Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 147

23. MAGNETIC FABRICS AND SOURCES OF MAGNETIC SUSCEPTIBILITYIN LOWER CRUSTAL AND UPPER MANTLE ROCKS FROM HESS DEEP1

Carl Richter,2 Paul R. Kelso,3 and Christopher J. MacLeod4

ABSTRACT

We present a magnetic fabric study of mafic and ultramafic rocks recovered at Ocean Drilling Program Sites 894 and 895from Hess Deep, a tectonic rift in the equatorial East Pacific (2°15'N, 101°30'W). We demonstrate, using thermomagneticcurves and high-field (up to 1.5 T) susceptibility measurements, that the magnetite contributes >95% of magnetic susceptibilityof these rocks. Bulk susceptibilities for the Site 894 gabbroic rocks are on average 1.72 × I0"2 and for the Site 895 peridotites4.29 × 10~2 (S1 volume units). The anisotropy of magnetic susceptibility is moderate to high and average kmw/kmjn is 1.15 ikmax

kint kmin are the principal susceptibilities) at Site 894 and 1.11 at Site 895. Because of the pervasive serpentinization and theformation of randomly distributed secondary magnetite, the magnetic fabrics of the Site 895 peridotites show no apparent rela-tionship to structural features. In the foliated gabbroic rocks from Site 894, we observe a close relationship between magmaticflow fabrics (defined by the preferred orientation of Plagioclase) and magnetic fabrics: kmin is perpendicular to the macroscopicfoliation and kmax dips steeply within the foliation plane and is parallel to the magmatic mineral lineation. Macroscopically iso-tropic gabbros yield the same principal susceptibility directions as the foliated gabbros. We interpret magnetic fabrics in theserocks to represent a weakly developed rock fabrics. We argue that the process for the development of the AMS is a distributionanisotropy which is caused by the growth of equant, irregular, or skeletal magnetite grains into a preferredly oriented (by mag-matic flow) Plagioclase "template." Magnetic fabric data record a mineral preferred orientation with a north-south, East PacificRise parallel strike and a near-vertical inclination, and can be interpreted to record the upward flow of melt at the top of an axialmagma chamber into the base of an overlying sheeted dike complex.

INTRODUCTION

Sites 894 and 895 of Ocean Drilling Program (ODP) Leg 147 arelocated in Hess Deep on the eastern flank of the East Pacific Rise(EPR) in the equatorial East Pacific (Fig. 1). The unique position atthe tip of a westward-propagating rift between the Cocos and NazcaPlates makes Hess Deep an ideal target to study magnetic propertiesin a vertical section of lower crustal and upper mantle rocks at a fastspreading (130 mm/yr) oceanic ridge. The western end of the rift val-ley is situated in approximately 1-Ma-old EPR crust. The drill sitesare located on an intrarift ridge in water depths of about 3023 m (Site894) and about 3820 m (Site 895) in an area that has been geological-ly mapped by submersibles (Francheteau et al, 1990; Karson et al.,1992).

We discuss the rock magnetic properties of basically two differentrock types (peridotites and gabbroic rocks, that are both affected byvarying degrees of serpentinization and alteration). Hole 894G wasdrilled to 154 meters below seafloor (mbsf) and mainly gabbronoritesand olivine gabbros were recovered; Hole 895D was drilled to 94mbsf and mainly harzburgites were recovered. Hole 895E reached 88mbsf and was drilled mainly through dunites, gabbroic rocks, andharzburgites (Gillis, Mével, Allan, et al., 1993).

894.^95

Hess Deep

Figure 1. Location of Leg 147 Sites 894 and 895 and generalized tectonicsetting of the Hess Deep rift in the equatorial East Pacific (after Lonsdale,1988) near the junction of the East Pacific Rise (EPR) and the Galapagosspreading center (inset).

'Mével, C , Gillis, K.M., Allan, J.F., and Meyer, P.S. (Eds.), 1996. Proc. ODP, Sci.Results, 147: College Station, TX (Ocean Drilling Program).

2 Ocean Drilling Program and Department of Geology and Geophysics, Texas A&MUniversity, 1000 Discovery Drive, College Station, TX 77840, U.S.A.

'Institute for Rock Magnetism, University of Minnesota, Minneapolis, MN 55455.(Current address: Lake Superior State University, Sault Sainte Marie, MI 49783,U.S.A.)

"Department of Earth Sciences, University of Wales College of Cardiff, P.O. Box914, Cardiff CF1 3YE, United Kingdom. Formerly at: Institute of Oceanographic Sci-ences, Brook Road, Wormley, Surrey GU8 5UB, United Kingdom; and BoreholeResearch, Department of Geology, University of Leicester, Leicester LEI 7RH, UnitedKingdom, macleod@cardiff.ac.uk

The anisotropy of magnetic susceptibility (AMS; reviews by Bor-radaile, 1988; Tarling and Hrouda, 1993) of most rock-forming min-erals (e.g., paramagnetic muscovite, biotite, chlorite) is linked to thecrystallographic lattice, that is, the principal susceptibility axes de-pend on the crystal symmetry (Nye, 1985), parallel to the crystallo-graphic axes (magnetocrystalline anisotropy; e.g., Borradaile andWerner, 1994). This relationship allows the use of magnetic anisotro-pies as an easy-to-apply proxy for measuring crystallographic pre-ferred orientations (Hrouda and Schulman, 1990; Richter et al., 1993)if it can be demonstrated that the magnetic fabric is due to minerals

C. RICHTER, P.R. KELSO, C.J. MACLEOD

with a magnetocrystalline anisotropy (e.g., Housen and van der Plu-ijm, 1990). The magnetic anisotropy of ferrimagnetic magnetite iscaused by shape anisotropy (Uyeda et al., 1963); that is, for multido-main magnetite, the maximum susceptibility is parallel to the longmineral axis and the minimum susceptibility parallel to the short axis.The magnetic susceptibility of magnetite is many times higher thanthe susceptibility of paramagnetic minerals and small traces of mag-netite can entirely dominate the magnetic susceptibility. Alignmentor preferred distribution of (equant) magnetite grains can lead to adistribution anisotropy (Wolff et al., 1989; Hargraves et al., 1991;Stephenson, 1994). AMS data can only be interpreted meaningful ifthe sources of magnetic susceptibility are known.

The anisotropy of remanence, especially the anisotropy of anhys-teretic remanent magnetization (AARM; McCabe et al., 1985; Jack-son, 1991), measures the anisotropy of ferro(i)magnetic mineralsonly. In this technique, an anhysteretic remanent magnetization(ARM) is imparted to a sample using a low steady (bias) magneticfield superimposed on the alternating field of an alternating field(AF) demagnetizer. The resulting ARM is measured and the processis repeated in nine different orientations. If both the AF and the directcurrent (DC) field are well below the coercivities of hematite, goe-thite, and fine-grained pyrrhotite, the AARM can only be due to themagnetic anisotropy of magnetite, coarse-grained pyrrhotite, orgreigite.

The few studies on the magnetic anisotropy of ultramafic rocksfrom the oceanic crust (Smith and Banerjee, 1985; MacDonald andEllwood, 1988; Bina et al., 1990; Bina and Henry, 1990) have shownthe existence of relatively strong magnetic anisotropies. Serpenti-nized peridotites from Hole 670A in the Mid-Atlantic Ridge south ofthe Kane Fracture Zone show that kmin clusters around [010] of oliv-ine (i.e., parallel to the spinel foliation) and kmax around the spinel lin-eation (Bina et al., 1990). This relationship appears to be invalid forhighly serpentinized samples. Magnetic fabrics of samples from theKane Fracture Zone are often related to the high-temperature defor-mation developed prior to serpentinization: the maximum axes thatcorrespond to the long axes of magnetite particles are parallel to thetectonic lineation, and the minimum axes (perpendicular to the flat-tening of magnetite particles) are perpendicular to the foliation. Ex-tensively serpentinized samples contain more dispersed secondarypseudo-single-domain magnetite and show little or no relationshipbetween olivine crystallographic preferred orientation (CPO) andmagnetic fabric. Wagner et al. (1981) studied the magnetic anisotro-py of ophiolitic gabbros from the western Alps. The observed suscep-tibilities are low (200-578 × 10~6) and the magnetic fabric isdominated by paramagnetic hornblende and augite minerals. In thiscase, AMS determines the CPO of the paramagnetic minerals, wherekmjn is parallel to the pole of the foliation plane and kmax parallel to thetectonic lineation. Ernst and Baragar (1992) inferred the magma flowpattern and magmatic processes of a huge dike swarm from AMSmeasurements on mafic igneous rocks which are macroscopicallyisotropic.

AMS is a very sensitive petrofabric tool, especially in rocks thatlack a macroscopic fabric, as is the case in many of the isotropic-looking gabbros recovered at Hess Deep. Using 138 specimens ob-tained during Leg 147, our goal was to (1) quantify the sources ofmagnetic susceptibility and the origin of magnetic anisotropy in thesemafic and ultramafic rocks, (2) investigate the orientation and shapeof induced and remanent magnetic anisotropies, (3) use magneticanisotropies to characterize the rock fabric, and (4) relate magneticanisotropies to observed structural features.

EXPERIMENTAL PROCEDURES

Cylindrical specimens (diameter = 2.5 cm, length = 2.2 cm) weredrilled from the working half of sampled cores. The anisotropy of

low-field magnetic susceptibility (AMS) of the specimens was mea-sured using a KLY-2 Kappabridge (Geofyzika Brno, inducing field =0.5 mT) on the JOIDES Resolution and with a custom-built suscepti-bility bridge (inducing field = 0.1 and 1.0 mT) at the Institute forRock Magnetism (IRM) at the University of Minnesota. The CS-2furnace attachment for the Kappabridge at the IRM was used to de-termine the temperature dependence of magnetic susceptibility be-tween room temperature and 700°C. For AARM analysis, specimenswere first demagnetized using a Schonstedt GSD-1 alternating field(AF) demagnetizer to obtain a baseline measurement that accountsfor the remanent magnetizations that the AF demagnetizer did not re-move. A DC-powered coil around the AF demagnetizer coil gener-ates a steady magnetic field of 0.05 mT. The ARM was generatedover the 0-95 mT field range in nine different orientations. The sam-ple was demagnetized after each step. The AARM tensor was calcu-lated following the method of Girdler (1961). High-fieldsusceptibilities (up to 1.5 T) were determined using the Princeton Ap-plied Research vibrating sample magnetometer (VSM) at the IRM.Magnetic anisotropy measurements are presented in Table 1.

MAGNETIC SUSCEPTIBILITY

The magnetic susceptibility (&,-,) is the dimensionless proportion-ality factor between the magnitude of induced magnetic intensity (M,)and the applied magnetic field strength (HJ):

, 1985),

where ki} is a second-rank symmetric tensor that can be visualized byan ellipsoid. The principal axes of the magnetic susceptibility ellip-soid are kmax kint kmin (e.g., Hrouda, 1982). Bulk susceptibility isdefined as

The ratio

= (kmax + kim + kmin)/3 (Nagata, 1961).

, 1961)

is used to quantify the degree of anisotropy. S1 volume units areused throughout.

The ellipsoids are represented in two types of diagrams that plotthe principal values (directions and magnitudes): (1) lower hemi-sphere equal area stereograms of the principal directions using theconventions proposed by Ellwood et al. (1988) and (2) Flinn-type di-agrams (Flinn, 1962) of the ellipsoid shape in the two-dimensionalspace.

SOURCES OF MAGNETIC SUSCEPTIBILITY

We applied two methods to investigate the sources of magneticsusceptibility: (1) the comparison between high-field (kHF) and low-field (kLF) susceptibility (Rochette and Fillion, 1988; Hrouda and Je-linek, 1990), and (2) thermomagnetic behavior (e.g., Bina and Henry,1990; Schmidt, 1993) using k vs. temperature k(T) curves. A discus-sion and comparison of both methods is given in Richter and van derPluijm (1994). In high magnetic fields (1-1.5 T), above the saturationmagnetization of the ferrimagnetic minerals, only diamagnetic (kD),paramagnetic (kP), and antiferromagnetic (kAF) susceptibilities aremeasured: kHF = kD + kP + kAF. In low fields (0.1-0.5 mT) the ferri-magnetic minerals (kF) also contribute to the magnetic susceptibility,so that kLF - kF + kD + kP + kAF. It follows from these equations that

394

MAGNETIC FABRICS IN ROCKS FROM HESS DEEP

the ferrimagnetic contribution to magnetic susceptibility is calculatedby

kp — kLF - kHF.

Acquisition curves of the isothermal remanent magnetization(IRM) of samples from both sites show that magnetic saturation oc-curs well below 200 mT (Fig. 2), which is indicative of magnetite ormaghemite. The differences in the saturation magnetization are dueto the magnetite concentration. At both sites, pseudo-single-domainmagnetite is the magnetic carrier mineral (for a detailed characteriza-tion of the magnetic mineralogy, see Kelso et al., Pariso et al., boththis volume). At both sites, kHF is between one and two orders of mag-nitude lower than kLF (Fig. 3A, B) which demonstrates that kF domi-nates the low-field magnetic susceptibility. Figure 3C shows 100[(A:iF- kHF)lkLF]% for Site 894 and Site 895 samples plotted vs. bulk sus-ceptibility. The graph shows that the ferrimagnetic dominance in-creases with increasing bulk susceptibility which implies that thebulk susceptibility is a function of the magnetite concentration. Thehistogram of the bulk susceptibilities (Fig. 3C), in comparison withthe percent ferrimagnetic contribution, demonstrates that the suscep-tibility of all samples is strongly magnetite-dominated.

Figure 4 shows two k(T) curves (25°-700°C) for two samplesfrom Sites 894 and 895. The arrow indicates the heating and the cool-ing parts of the curve. Both curves are essentially reversible, indicat-ing little chemical change during the heating process. Sharp Curietemperatures (Tc) at 562°C (Sample 147-894G-20R-3, 124 cm) and

Site 894

CO

c

Site 895

200 400 600

Applied field [mT]

800

Figure 2. Acquisition curves of the isothermal remanent magnetization. Mag-netic saturation for all samples occurs well below 200 mT, indicating magne-tite or maghemite.

575°C (Sample 147-895E-1R-2, 125 cm) indicate the presence ofmagnetite with low Ti content and almost pure magnetite (Tc of puremagnetite = 580°C) (see Kelso et al., Pariso et al., both this volume).Below Tc, the magnetic susceptibility stays constant or increasesslightly with increasing temperature. This behavior is typical for thethermomagnetic behavior of magnetite (e.g., Schmidt, 1993) withouta paramagnetic contribution. A very shallow Hopkinson peak can beobserved in both samples. A susceptibility decrease with increasingtemperature following the Curie-Weiss law would indicate an influ-ence of paramagnetic material to the bulk susceptibility (Schultz-Krutisch and Heller, 1985; Richter and van der Pluijm, 1994).

Both methods demonstrate that the dominant source of magneticsusceptibility and, hence, of magnetic susceptibility anisotropy ismagnetite for both the Site 894 gabbros and gabbronorites and theSite 895 peridotites. Magnetic fabrics, therefore, represent the overallshape and distribution of magnetite in these rocks. Olivine, pyroxene,amphiboles, serpentine, ilmenite, pyrrhotite, and the diamagnetic pla-gioclase also contribute to some degree, but have only very minor in-fluence on the magnetic susceptibility tensor.

BULK SUSCEPTIBILITY

Figure 5 shows the observed magnetic bulk susceptibilities sepa-rately for the three principal holes. The values are high and rangefrom 2.78 × 10^ to 1.35 x K H The mean value for the Site 894 rocksis 1.72 × I0"2 and for the Site 895 peridotites, 4.29 × I0'2, which issimilar to previously reported values from oceanic mafic and ultra-mafic rocks (Fox and Opdyke, 1973; Kent et al., 1978). Susceptibilityvalues between 4.27 × 10"* and 4.32 × I0"1 have been determinedfrom the ODP Hole 735B gabbros (Kikawa and Pariso, 1991).

The vertical distribution of magnetic susceptibilities in Holes894G, 895D, and 895E is shown in Figure 5. Magnetite is the majorsource of magnetic susceptibility and, hence, variations in the mag-netic susceptibility directly reflect variations in the concentration ofmagnetite (Kelso et al., this volume). Hole 894G shows no systematicdistribution of magnetic susceptibilities with depth. A correlation be-tween Lithology (see simplified lithostratigraphic column in Fig. 5)and k does not exist. Hole 895D shows a trend towards higher suscep-tibilities downhole, which indicates an increase in the concentrationof magnetite. Again, no relationship between k and lithology is evi-dent. Samples from Hole 895E reach the highest observed suscepti-bilities and display a large scatter in bulk susceptibilities.

MAGNETIC FABRICS

Anisotropy of Magnetic Susceptibility

Ferrimagnetic minerals have a shape anisotropy that is caused bythe demagnetizing field, Hdemag, which reduces the external field Hext

depending on the shape and orientation of the magnetized body. Theeffective field HeJfis

Heff=(\lk~N)M

where ./Vis the demagnetizing factor (e.g., Osborn, 1945). Pfleidererand Halls (1990) investigated the relationship between shape andAMS of artificial samples and found that AMS is related inversely tothe demagnetization factor, which itself relates inversely to the axialratios of the ferrimagnetic grains. AMS provides a direct measure ofthe overall magnetite shape. A second mechanism to obtain a signif-icant ferrimagnetic anisotropy is the distribution of the grains. Forexample, an alignment of spherical magnetite grains in a plane willlead to a higher susceptibility parallel to that plane than perpendicu-lar to it. This phenomenon has been described as distribution anisot-ropy (Hargraves et al., 1991).

395

C. RICHTER, P.R. KELSO, CJ. MACLEOD

ß 0.002

0.0015SCD

g> 0.001

0.0005

Hole 894G

Ferrimagnetic: 97%

"-J"

B

field

)'

(hig

h

0.0015

0.001

0.0005

n

Site 895

Ferrimagnetic:

•

• •

L•

97%

•

•

0.01 0.02 0.03 0.04 0.05

k (low field)

0.05 0.1

k (low field)0.15

100

0.05 0.1k (low field)

Figure 3. A, B. High-field susceptibility plotted vs. low-field susceptibility. The low-field susceptibility is 1-2 orders of magnitude higher, indicating a strongmagnetite dominance. The contribution of ferrimagnetic magnetite to the low-field susceptibility is on average 97% at both sites. C. Ferrimagnetic contribution(in percent) to low-field susceptibility of individual samples plotted vs. bulk susceptibility and histogram of bulk susceptibilities. Most samples have a suscepti-bility that is dominated by magnetite to at least 95%.

AMS results are presented in Table 1. The magnetic anisotropy ismoderate to high in most samples. Figure 6 shows the degree ofanisotropy P plotted vs. depth and lithology. Anisotropies are highestin Hole 894G gabbros and there is a tendency of increasing anisotro-py downhole to about 100 mbsf, where the highest anisotropies arereached. P varies between 1.01 and 1.42 (average = 1.15) at Hole894G, between 1.03 and 1.18 (average =1.10) at Hole 895D, and be-tween 1.03 and 1.25 (average = 1.12) at Hole 895E. The lithology hasno apparent influence on the anisotropy or on the ellipsoid shape.AMS ellipsoid shapes are presented in Flinn diagrams (Flinn, 1962)in Figure 7. The shapes of the susceptibility ellipsoids scatter be-tween oblate, prolate, and almost isotropic. Most are triaxial, that is,they cluster around the dashed line of the Flinn diagrams (kmajkint =kjnjkmin), except that samples from Hole 895E tend toward more ob-late-shaped AMS ellipsoid forms.

The rotary-drilled cores have only a vertical (no horizontal) orien-tation and the principal susceptibility directions appear to be randomfor that reason (Table 1). In addition, the complicated tectonic situa-tion within the rift (e.g., Karson et al., 1992) adds another factor thatapparently randomizes the orientation of the susceptibility ellipsoids.Stable paleomagnetic directions have an average inclination of 38° atSite 894 (Pariso et al., this volume) and no consistent inclination atSite 895 (Kelso et al., this volume), although the calculated dipolefield inclination at the latitude of Hess Deep is +4.6°. This impliesthat major block rotations occurred, and not only the horizontal ori-entation but also the vertical orientation before tectonic rotation anddrilling is uncertain. The dip of the principal susceptibility axes, how-ever, should be unaffected by the lack of azimuthal control andshould yield an estimate of the mean inclination of the susceptibility

axes after the rift-related tectonic rotation. Figure 8 shows histogramsof the dip of kmin for the three principal holes. At Hole 894G it is ap-parent that the pattern is not random and that most kmin axes have ashallow dip with an average of 25°. The kmin distribution at Hole895D is very scattered and shows no significant maximum. The scat-ter is also considerable at Hole 895E but a distinct maximum isreached at 30°. The dip of kmin is generally shallow and shows a verysystematic distribution at Holes 894G and 895E. This suggests thatthe orientation of the AMS ellipsoids in a geographical referenceframe bear a geological significance.

Anisotropy of Anhysteretic Remanent Magnetization

ARM anisotropy measures a remanent magnetization and, hence,determines the fabric of the carriers of remanence, that is, magnetiteonly (Kelso et al., Pariso et al., both this volume). We have demon-strated that AMS is also dominated by magnetite. Both methodsshould yield a comparable result if the measured effective magneticgrain size is the same. Figure 9 compares the AMS and the AARMellipsoids of three samples from the three main holes. Figure 9Ashows the magnetic fabrics of a gabbronorite from Hole 894G. Theshape of the prolate ellipsoids and the orientations of the principal di-rections are virtually the same, but PAMS is higher than PAARM Aharzburgite sample from Hole 895D (Fig. 9B) has triaxial ellipsoidshapes and the principal directions of the AMS and AARM ellipsoidsare only a few degrees apart. The dunite sample from Hole 895E (Fig.9C) has a strongly oblate fabric, / ‰ M is 1.38 and very strong com-pared to PAMS = 1.02, which is typical of the methods (Jackson, 1991).The orientation of the minimum axes is virtually identical, whereas

MAGNETIC FABRICS IN ROCKS FROM HESS DEEP

the maximum and intermediate directions are rotated, as can be ex-pected for a strongly oblate ellipsoid, on a great circle peΦendicularto kmin. Reasons for the deviation between the two tensors are that (1)AMS, though it is dominated by magnetite, is also influenced byparamagnetic and diamagnetic susceptibilities, and (2) AARM andAMS do not measure the same effective magnetic grain size.

DISCUSSION

The orientation of the magnetic fabric of ultramafic rocks hasbeen demonstrated to be related to structural features and flow pat-terns (Smith and Banerjee, 1985; MacDonald and Ellwood, 1988;Bina et al., 1990; Wagner et al., 1981). The macroscopic fabrics ofthe gabbroic rocks collected at Site 894 are either isotropic or foliat-ed. The foliation is defined by the shape-preferred orientation of crys-tals, mainly Plagioclase, and is related to magmatic flow (Gillis,Mével, Allan, et al., 1993). In contrast to samples from Hole 735B(ODP Leg 118), the rocks show little evidence for solid-state pene-trative deformation. The orientation of the near-vertical foliation isgenerally north-south with a steeply plunging lineation (Fig. 10;MacLeod, Boudier, et al., this volume).

Figure 11 shows the principal susceptibility axes of foliated gab-bros from Hole 894G. The data are corrected for the stable magneticremanence (Kelso et al., this volume; Pariso et al., this volume), thatis, to a common declination (0°, normal polarity) and inclination(+4.6°, the calculated dipole field inclination for the latitude of HessDeep). A comparison with the structural data from Figure 10 showsthat kmax is parallel to the mineral lineation and that kmin is parallel tothe poles of foliation. kmax has a steeply dipping orientation and scat-ters on a great circle within the foliation plane. kint and kmin are shal-low-dipping. The magnetic fabric is directly related to the rock fabric,though the rock fabric is defined by the shape-preferred orientationof mainly Plagioclase, which does not significantly contribute toAMS.

We have used this relationship to investigate the fabrics of mac-roscopically isotropic-looking gabbro samples from Hole 894G (Fig.12). Compared to the fabric diagrams (Fig. 10) and the magnetic fab-

CO

cd

Sample 147-894G-20R-3,124 cm

Sample 147-895E-1R-2,125cm

~100 200 300 400 500 600 700 800Temperature (°C)

Figure 4. Heating and cooling curves of low-field susceptibility. Arrowsindicate heating and cooling part of curve. No temperature dependence of thesusceptibility before the intensity breakdown at the Curie temperature indi-cates that only ferrimagnetic magnetite contributes significantly to magneticsusceptibility. Both curves are virtually reversible, indicating little chemicalchange during heating. The Hopkinson peak is very shallow in both samples.

Hole 894Gk

0.04 0.08

Hole 895Dk

0.04 0.08

Hole 895Ek

0.05 0.1 0.15

90

Lithology131 Olivin gabbroE3 01. gabbro norite• l Basalt dikeCM Gabbronoriteβ Gabbro

E3 HarzburgiteD i DuniteC2l Troctoliteg3 Dunite/harzburgite/

troctolite/gabbro

150

Figure 5. Bulk susceptibility plotted vs. depth andsimplified lithology. The susceptibility magnitude isdirectly related to magnetite concentration.

397

C. RICHTER, P.R. KELSO. C.J. M A C L E O D

Table 1. Anisotropy of magnetic susceptibility.

Core, section,

interval (cm)

147-894G-1R-1, 172R-1.682R-2, 572R-2, 1102R-3, 322R-3, 1183R-1,64R-2, 766R-1, 266R-1,776R-1.896R-2, 06R-2, 787R-1,997R-2, 67R-2, 107R-2, 618R-1, 548R-2, 319R-1, 1209R-2, 869R-3, 199R-3, 699R-3, 779R-4, 459R-4, 909R-4, 959R-4, 130lOR-1,8110R-l,9110R-2, 8HR-2,3411R-2, 4512R-2, 5012R-2, 11112R-2, 14012R-3, 412R-3, 4812R-3, 6712R-4, 2012R-4, 5112R-4, 13912R-5.7912R-6, 3913R-1, 5913R-1.6213R-2, 12013R-3, 513R-3, 3014R-1.7015R-1.73ΠR-I,60

ΠR-1,90

17R-1, 12317R-2, 11418R-1.6918R-2, 718R-2, 9420R-1.6920R-3, 5620R-3, 124

147-895D-2R-1, 252R-1.472R-2, 392R-2, 512R-2, 783R-1.454R-1.414R-3, 44R-3, 104R-3, 154R-4, 844R-4, 1004R-5, 34R-5, 85R-1,525R-2, 1006R-1.326R-1.757R-1.657R-1, 1197R-2, 137R-2, 927R-2, 1078R-1.458R-2, 4

Depth

(mbsf)

18.7729.2831.0031.2031.8732.7338.2647.1055.0655.5755.6956.2357.0165.7966.2666.3066.8169.0470.3175.6076.5177.1077.6077.6878.7878.4279.2879.6379.9180.0180.5685.9086.0195.7696.3796.6696.7797.2197.4098.4398.7499.62100.52101.56103.99104.02105.97106.28106.53113.80119.53126.40126.70127.03128.38131.59132.39133.26146.20147.98148.66

16.2016.4717.8117.9318.2026.4535.1137.0338.0937.1439.2939.4539.7639.8143.8245.7455.3255.7565.2565.7966.2066.9967.1474.7575.72

0.0096820.0071910.0081540.0085760.0019090.0257270.0077650.0036800.0071170.0039830.0130020.0737230.0182970.0059770.0109130.0104260.0046720.0128940.0107380.0149100.0147330.0189610.0331910.0347410.0079850.0048990.0105430.0163110.0164550.0030370.0121090.0042480.0075410.0262850.0305090.0225970.0114770.0098860.0078040.0148890.0185860.0120380.0567040.0425750.0152570.0140490.0186110.0405470.0155300.0111360.0171410.0081810.0133770.0107490.0249270.0044160.0043080.0064440.0119300.0207640.006438

0.0058990.0101840.0239730.0337890.0198020.0054150.0036900.0102060.0024100.0137980.0033750.0077050.0046550.0034300.0230740.0109610.0003430.0312610.0200260.0269650.0301710.0293630.0194910.0276320.000304

D

3542662051552402761143431818712511135684

109136280195257161403012583302282173385650174289207047300302662520

14234822512311229017131

581583421161061

7510418116

270

10423097103103222171156233285297182534629421931730630029315135035275342

I

7131736391291132560385611815222015S2213222373435542529574512143540152032428224463460I014637241521542314267197921

58231753133592035472746546214034262482811710

kin,

0.0098720.0072420.0082010.0086170.0019250.0264940.0081100.0037490.0072580.0040120.0134650.0828860.0194560.0065150.0121570.0109640.0049480.0132550.0118090.0157410.0162480.0203850.0359200.0365260.0086260.0053260.0110930.0169990.0167860.0030880.0135030.0044420.0084170.0351950.0323320.0241850.0125520.0110080.0079480.0162590.0188740.0126850.0635210.0532270.0196070.0147900.0197050.0450190.0186790.0121300.0182980.0085340.0153800.0111280.0259400.0049780.0045080.0064810.0122300.0215360.006599

0.0062390.0103380.0265050.0371730.0201390.0059870.0037960.0107930.0024780.0141420.0034870.0080050.0048680.0035740.0249210.0115700.0003470.0341790.0209800.0279450.0312810.0334030.0221110.0286650.000316

D

10311111027

35117442253280196347129361629421430273979

206502083501263289524232526716165283297139174291360346269242803282633193126413623160302843342693073491783485226142

215230347348197129272563393260891782071583125857848330418713033984

I

717617402444

532363523315830183774662154824651153742563129317033530961618728365261445351561386474216237462561

768385432651625939374311151922523506164822049

kmajc

0.0100010.0072730.0082940.0086430.0019620.0271590.0081550.0039100.0076750.0042340.0152030.0921050.0208970.0068480.0133200.0113520.0051180.0137210.0121330.0165420.0166890.0220160.0397590.0434810.0090180.0058030.0113380.0176300.0172490.0031410.0141330.0048610.0091720.0363160.0332500.0271790.0136300.0119550.0089550.0190960.0209800.0139070.0710840.0585400.0216250.0154420.0206900.0504840.0211660.0125140.0183840.0089820.0174710.0114190.0278030.0052310.0047330.0068330.0128600.0222120.006729

0.0064980.0105170.0276870.0386640.0211400.0060550.0039230.0115790.0025280.0151870.0037040.0082100.0050090.0036270.0260790.0118310.0003540.0351670.0223340.0288580.0325770.0346730.0228280.0292960.000319

D

26235733726910518

241967532924924522832826935945186325158116314112952317833514923015028056129164234591431138510336018810110

2111254628819026861235212174171212303195278170172

133342112035292736413717418027727129862441921691971975697220203243

I

18666264143

35

514418373160714741061321

69576712523435651710315614543156675773566952211329734154491548143538695315201020

13624331575818

294433471551644564119131121

MAGNETIC FABRICS IN ROCKS FROM HESS DEEP

Table 1 (continued).

Core, section,interval (cm)

9R-1, 129R-1,789R-1, 1299R-1, 140

147-895E-lR-1,471R-2, 721R-2, 1051R-2, 1251R-3, 241R-3, 491R-3, 1222R-2, 562R-2, 1073R-2, 243R-2, 553R-2, 1363R-3, 144R-1.924R-1, 1024R-2, 464R-2, 584R-3, 65R-1, 1455R-2, 485R-2, 1155R-2, 1325R-3, 356R-1, 1316R-2, 216R-2, 1116R-3, 686R-3, 1016R-5, 406R-5, 666R-5, 746R-5, 957R-1, 1357R-2, 107R-3, 397R-3, 607R-3, 1257R-4, 368R-1, 1138R-1, 1208R-2, 168R-2, 578R-2, 708R-2, 908R-3, 518R-3, 668R-3, 1038R-4, 118

Depth

(mbsf)

85.2084.7885.2985.40

0.472.162.492.693.103.354.0821.5222.0331.2131.5232.3332.6140.4240.5241.4041.5242.4750.3550.8851.5551.7252.2559.9160.3161.2161.9962.3264.1564.4164.4964.7069.5570.7271.5971.8072.4573.0679.0379.1079.5579.6980.0980.2981.3181.4681.8383.48

Kin

0.0624380.0433140.0558460.068155

0.0003150.0728910.0631940.0907230.0558290.1020410.0004300.0742100.0551420.0723600.1020660.1229540.1137590.0417340.0563300.0476800.0494310.0002750.0645300.0512760.0686380.0509010.0462350.0470980.0383640.0260740.0235870.0239270.0424730.0446200.0440800.0441580.0558260.0432510.0352400.0337800.0570500.0120200.0426070.0441560.0676310.0580250.0361130.0352800.0474640.0604820.0652030.061090

D

71313300174

666010785512892842027126924432431860561703314619

23338492441862521102532981633333203001801001663391514584831305519859104185114229

I

2350186

482350436073485033153053681627172423236701022233533402526101047715645494039283530243145285118

Kn,

0.0687650.0438290.0616050.070881

0.0003230.0746410.0704920.0948690.0598190.1176970.0004460.0754200.0592780.0807610.1211790.1395500.1277660.0451260.0586390.0511470.0521560.0002770.0666820.0537130.0792830.0567160.0515010.0497420.0413570.0276890.0254160.0258290.0469320.0465530.0482910.0453450.0619590.0466700.0381120.0354720.0628490.0127480.0458190.0479300.0745190.0589130.0382670.0373360.0498220.0732450.0694710.063773

D

327765958

1673212232472479419115461139589818328926560218142111360196178111734529032

2568251343137

2591092546

3382532471723391572572813453

I

30

5677

102221462916330536760281667

48411334401974

4352574262663537811233244219623339601342102765

kmax

0.0704930.0479160.0629420.074283

0.0003250.0759230.0718590.0994980.0641880.1206240.0004540.0810200.0651810.0830090.1230030.1436620.1346630.0518140.0603510.0513680.0540880.0002830.0711410.0555160.0810000.0573320.0523800.0509290.0439450.0279060.0259030.0266540.0479080.0482490.0491660.0478180.0626970.0494200.0404180.0376680.0645260.0134520.0474860.0512630.0793020.0606830.0390920.0391860.0509110.0757800.0705590.064851

D

192182200265

265192327347153185982591714

152200891551532738226028611828931734229515220142202359238168205902773502181482542283517

29910026636029240133

I

51292811

4157339744224151732315171237396356307132038130221264257953719342731224543737175514602617

Notes: AMS measurements in S1 units (volume susceptiblity). D = declination in degrees; I = inclination in degrees (core coordinates).

rics of the foliated gabbros (Fig. 11), the magnetic fabrics of the un-foliated samples show the same characteristics. kmax dips steeply tothe northeast. The distribution of kmin and kint is more scattered. Mostkmin axes have an east to northeast direction and most kim axes dipshallowly to the south. Though a macroscopic fabric is not visible, wecan infer from the magnetic fabric data that the isotropic gabbroshave a steeply dipping magmatic lineation and north-south-orientedfoliation. These structures are too weakly developed to be visible butcan be detected with sensitive petrofabric indicators, like AMS.

We argue, that the process for the development of the AMS is adistribution anisotropy (Wolff et al., 1989; Hargraves et al., 1991;Stephenson, 1994). Magnetic grains do not have to be nonequant (el-liptical, elongated, or flattened) and aligned in order to impart a de-tectable AMS. The effect may arise through organization of equant,irregular, or skeletal grains into planar or linear arrangements. It isnot necessary that the grains actually grow during the magma flowperiod (and thus acquire a shape anisotropy), it is enough that thegrains are arranged or grow along preexisiting fabric elements. Har-graves et al. (1991) argue that AMS in pristine igneous rocks is a di-rect or indirect reflection of preexisting silicate fabric, which isassociated with flow or intrusion of the magma. They conclude that

AMS provides a direct record of fluid-dynamic histories of igneousrocks. The Hole 894G plutonic rocks contain primary igneous mag-netite and ilmenite, which are now affected by exsolution and alter-ation. Thin sections show that the texture of the oxide minerals ispredominantly interstitial, and equant in coarser and less deformedpatches (Gillis, Mével, Allan, et al., 1993). The late-crystallizing in-terstitial minerals form thin cuspate-shaped grains between the sili-cates, partially or totally enclose silicates, or, where interstitialbetween Plagioclase, make up subangular cross sections. The distri-bution of these grains will be relatively anisotropic if they grew in re-sidual liquid volumes within a preferredly oriented (by magmaticflow) silicate (i.e., a Plagioclase "template"). This mechanism ex-plains (1) the origin of magnetic anisotropy in mafic and ultra maficrocks from Hess Deep and (2) why macroscopic Plagioclase fabricsand magnetite fabrics caused by interstitial and sometimes equantgrains are parallel.

Magnetic fabric data demonstrate that the gabbros (Figs. 11 and12) have a very uniform fabric with a steeply dipping magnetic/mag-matic lineation and a near-vertical foliation (the plane perpendicularto kmin). The average strike of the foliation is approximately north-south, that is, parallel to the strike of the EPR spreading axis. Sheeted

399

C. RICHTER, P.R. KELSO, CJ. MACLEOD

Figure 6. The degree of anisotropy P = kma/kmin plot-ted vs. depth and simplified lithology.

Hole 894Gp

1.1 1.2 1.3 1.4 1.5

Hole 895Dp

1.1 1.2 1.3 1

Hole 895Ep

1.1 1.2 1.3 1.4

Lithology

Olivin gabbro E23 HarzburgiteOl. gabbro norite

• i Basalt dikeEU GabbronoriteH Gabbro

• i DuniteEZ1 Troctoliteg-^ Dunite/harzburgite/

troctolite/gabbro

*.1.2

×cα

J 1.1

prolate

••

PfcJJE ^ • 1

*Hole894G

_ •• •

oblate •

1.1

prolate

1.2 1.3 1.4

1.05

1

Hole 895D

oblate

1 1.05 1.1 1.15 1.2 1.25

1.15

1.1

1.05

1 u

prolateHole 895E

" . - . - / * " . oblate

1 1.05 1.1 1.15 1.2 1.25Kint/Kmin

Figure 7. The shape of the magnetic susceptibility ellipsoid in Flinn dia-grams. Most ellipsoids are triaxial; samples from Hole 895E have moreoblate-shaped susceptibility ellipsoids.

8

6

4

2

n

1

Hole 895D

1 ill

CD

erCD

15

10

II !

Hole 895E

I I lib m20 40 60 80

in inclination (°)

Figure 8. Histograms of the inclination of the minimum susceptibility axes ofthe three principal holes. A. The dip of kmin of the Hole 894G rocks is shal-low, about 25° on average. B. Hole 895D shows a large kmin scatter and noapparent maximum. C. Hole 895E samples show a large scatter, but have adistinct maximum at 30°.

MAGNETIC FABRICS IN ROCKS FROM HESS DEEP

H o l e 8 9 4 G

lineations

/

/ Sample894G-9R-3, 77cm

1 1.2 1.4

.-,' Sample* 895D-4R-3, 4cm

1 1.2 1.4

/ Sample' 895E-6R-5,74cm

π . . •1 1.2 1.4

Figure 9. Comparison between AMS and AARM ellipsoids for selected sam-ples. Square = maximum; triangle = intermediate; circle = minimum direc-tion; open symbols = AMS; solid symbols = AARM. A, B. The principaldirections and shapes of both ellipsoids compare well. C. The differences inorientation and degree of anisotropy demonstrate that for this sample bothmethods measure a different effective magnetic grain size.

dikes measured on the north wall of Hess Deep during Alvin dives(Karson et al., 1992) are near-vertical and also have an EPR-parallelnorth-south orientation. The steep lineations and foliations in thegabbros parallel to those in the sheeted dike complex have been in-terpreted to record the upward flow of melt at the top of the axialmagma chamber into the base of a sheeted dike complex (Nicolas etal., 1988; MacLeod and Rothery, 1992).

Peridotites from Site 895 yielded no consistent stable paleomag-netic direction (Kelso et al., this volume). The dip of kmin in the unro-tated data is very scattered, and a relationship between macroscopicfeatures and AMS is not apparent. The randomly distributed second-ary magnetite that was formed during the serpentinization processpresumably randomizes the magnetic fabric and overprints anymeaningful relationships between tectonic structures and magneticfabrics.

CONCLUSIONS

Our magnetic fabric study of mafic and ultramafic rocks fromHess Deep demonstrates that AMS is a very sensitive method that canbe used to identify and characterize rock fabrics even in rocks that ap-pear to be macroscopically isotropic. In detail we demonstrated that:

1. High-field/low-field comparisons and thermomagnetic behav-ior show that magnetic susceptibilities (and, hence, magnetic fabrics)are dominated by magnetite.

2. Magnetic susceptibility anisotropy presumably originates froma distribution anisotropy which is caused by the growth of magnetitegrains into a preexisiting silicate flow fabric. The degree of anisotro-

N = 19 contoured at2 3 4 5 6 7 timesuniform distribution

H o l e 8 9 4 G

foliations

N = 24contoured at2 3 4 5 timesuniform distribution

Figure 10. Structural features of foliated gabbros from Hole 894G (fromMacLeod, Boudier, et al., this volume). A. Steeply dipping lineations. B.Magmatic foliations that are near-vertical and strike north-south, approxi-mately parallel to the EPR axis.

py is moderate to high (average kmjkmin = 1.15 at Hole 894G, 1.10 atHole 895D, and 1.12 at Hole 895E), susceptibility ellipsoids are ingeneral triaxial. Lithology has no apparent influence on the shape ofthe anisotropy ellipsoid.

3. Anisotropies of remanence (AARM) and of susceptibility(AMS) both measure the overall preferred orientation or shapeanisotropy of magnetite grains. The effective grain size determinedby both methods is different, however, and orientation and shape ofthe anisotropy ellipsoids differ for that reason.

4. The mean dip of the minimum susceptibility axes after tectonicrotation due to rifting is 25° for the Hole 894 gabbroic rocks.

5. We did not observe a relationship between structural featuresand magnetic fabrics for the Site 895 peridotites. The formation ofrandomly oriented secondary magnetite during the serpentinizationprocess obliterates any meaningful correlation between structuralfeatures and magnetic fabrics.

6. Rotation of the AMS directions of foliated gabbros from Site894 into a common stable paleomagnetic direction (declination andinclination) shows that kmin is perpendicular to the observed magmat-ic foliation and kmax parallel to the magmatic lineation. The lineationis dipping steeply to the northeast and the foliation has a north-south

401

C. RICHTER, P.R. KELSO, C.J. MACLEOD

max

N = 21

max

contoured at2 3 4 5 6 timesuniform distribution

in t

contoured at2 3 4 5 timesuniform distribution

N = 31

N = 21 contoured at2 3 4 timesuniform distribution

contoured at2 3 4 timesuniform distribution

m in

N = 31contoured at2 3 4 timesuniform distribution

N = 21 contoured at2 3 4 5 timesuniform distribution

Figure 11. Principal magnetic susceptibility axes of macroscopically foliatedgabbroic rocks from Hole 894G. Values are restored to a common stablemagnetic declination and inclination. kmαx has a steep inclination that is par-allel to the magmatic lineation and kmin a shallow inclination that is parallelto the poles to planes of the foliated gabbros (compare to Fig. 10).

Figure 12. Principal magnetic susceptibility axes of macroscopically isotro-pic gabbroic rocks from Hole 894G. Values are restored to a common stablemagnetic declination and inclination. The orientation of the axes is the sameas in the macroscopically foliated rocks and parallel to the lineation and foli-ation observed in the foliated gabbros.

strike that is parallel to the EPR. Rock fabrics are defined by the pre-ferred orientation of Plagioclase and magnetic fabrics determine thepreferred orientation/distribution of magnetite.

7. Macroscopically isotropic gabbros have the same magnetic fab-rics as the foliated gabbros. The minimum susceptibility axes are ori-ented about east-west and the maximum axes are near-vertical.

8. Structural and magnetic fabric data reveal a preferred orienta-tion with an EPR parallel strike and a steep inclination. The rocks

402

MAGNETIC FABRICS IN ROCKS FROM HESS DEEP

show little evidence for penetrative deformation and the structuresare interpreted to represent magmatic flow fabrics. These fabrics areparallel to structures determined in the sheeted dikes from the northwall of Hess Deep (Karson et al., 1992) and can be interpreted torecord the upward flow of melt at the top of the axial magma chamberinto the base of a sheeted dike complex.

ACKNOWLEDGMENTS

Rock magnetic measurements were carried out at the Institute forRock Magnetism (University of Minnesota). The IRM is funded bythe Keck Foundation, the National Science Foundation, and the Uni-versity of Minnesota. We thank Chris Hunt and Jim Marvin for theirsupport. The thermomagnetic curves could not have been includedwithout the help of Peat Solheid. We thank Peter Clift and reviewersBrooks Ellwood, Bernie Housen, Jamie Allan, and Shanna Collie forhelpful comments on our manuscript. Shipboard scientists F. Boudi-er, B. Célérier, L. Kennedy, E. Kikawa, and J. Pariso contributedstructural and paleomagnetic data. Research was supported by theDeutsche Forschungsgemeinschaft, grant RI576/1-2 (C.R.), USSSPgrant 147-20719b (P.K.), and the Natural Environment ResearchCouncil (C.J.M.).

REFERENCES

Bina, M.M., and Henry, B., 1990. Magnetic properties, opaque mineralogyand magnetic anisotropies of serpentinized peridotites from ODP Hole670A near the Mid-Atlantic Ridge. Phys. Earth Planet. Inter., 65:88-KB.

Bina, M.M., Henry, B., and Cannat, M., 1990. Magnetic anisotropy andsome other magnetic properties of serpentinized peridotites from ODPHole 670A. In Detrick, R., Honnorez, J., Bryan, W.B., Juteau, T., et al.Proc. ODP, Sci. Results, 106/109: College Station, TX (Ocean DrillingProgram), 263-267.

Borradaile, G.J., 1988. Magnetic susceptibility, petrofabrics and strain. Tec-tonophysics, 156:1-20.

Borradaile, G.J., and Werner, T., 1994. Magnetic anisotropy of some phyllo-silicates. Tectonophysics, 235:223-248.

Ellwood, B.B., Hrouda, F., and Wagner, J.-J., 1988. Symposia on magneticfabrics: introductory comments. Phys. Earth Planet. Inter., 51:249-252.

Ernst, R.E., and Baragar, W.R.A., 1992. Evidence from magnetic fabric forthe flow pattern of magma in the Mackenzie giant radiating dike swarm.Nature, 356:511-513.

Flinn, D., 1962. On folding during three-dimensional progressive deforma-tion. Q. J. Geol. Sot: London, 118:385^133.

Fox, PJ., and Opdyke, N.D., 1973. Geology of the oceanic crust: magneticproperties of oceanic rocks. J. Geophys. Res., 78:5139-5154.

Francheteau, J., Armijo, R., Cheminée, J.L., Hekinian, R., Lonsdale, P., andBlum, N., 1990. 1 Ma East Pacific Rise oceanic crust and uppermostmantle exposed by rifting in Hess Deep (equatorial Pacific Ocean). EarthPlanet. Sci. Lett., 101:281-295.

Gillis, K., Mével, C, Allan, J., et al., 1993. Proc. ODP, Init. Repts., 147:College Station, TX (Ocean Drilling Program).

Girdler, R.W., 1961. The measurement and computation of anisotropy ofmagnetic susceptibility of rocks. Geophys. J. R. Astron. Soc, 5:34-44.

Hargraves, R.B., Johnson, D., and Chan, C.Y., 1991. Distribution anisotropy:the cause of AMS in igneous rocks? Geophys. Res. Lett., 18:2193-2196.

Housen, B.A., and van der Pluijm, B.A., 1990. Chlorite control of correla-tions between strain and anisotropy of magnetic susceptibility. Phys.Earth Planet. Inter., 61:315-323.

Hrouda, F., 1982. Magnetic anisotropy of rocks and its application in geol-ogy and geophysics. Geophys. Surv., 5:37-82.

Hrouda, F., and Jelinek, V., 1990. Resolution of ferrimagnetic and paramag-netic anisotropies in rocks, using combined low-field and high-field mea-surements. Geophys. J. Int., 103:75-84.

Hrouda, F., and Schulman, K., 1990. Conversion of the magnetic susceptibil-ity tensor into the orientation tensor in some rocks. Phys. Earth Planet.Inter., 63:71-77.

Jackson, M., 1991. Anisotropy of magnetic remanence: a brief review ofmineralogical sources, physical origins, and geological applications, andcomparison with susceptibility anisotropy. Pure Appl. Geophys., 136:1-28.

Karson, J.A., Hurst, S.D., and Lonsdale, P.F., 1992. Tectonic rotations ofdikes in fast-spread oceanic crust exposed near Hess Deep. Geology,20:685-688.

Kent, D.V., Honnorez, B.M., Opdyke, N.D., and Fox, PJ., 1978. Magneticproperties of dredged oceanic gabbros and source of marine magneticanomalies. Geophys. J. R. Astron. Soc, 55:513-537.

Kikawa, E., and Pariso, J.E., 1991. Magnetic properties of gabbros fromHole 735B, Southwest Indian Ridge. In Von Herzen, R.P., Robinson,P.T., et al., Proc. ODP, Sci. Results, 118: College Station, TX (OceanDrilling Program), 285-307.

Lonsdale, P., 1988. Structural pattern of the Galapagos microplate and evolu-tion of the Galapagos triple junction. /. Geophys. Res., 93:13551-13574.

MacDonald, W.D., and Ellwood, B.B., 1988. Magnetic fabric of peridotitewith intersecting petrofabric surfaces, Tinaquillo, Venezuela. Phys.Earth Planet. Inter., 51:301-312.

MacLeod, C.J., and Rothery, D.A., 1992. Ridge axial segmentation in theOman ophiolite: evidence from along-strike variations in the sheeteddyke complex. In Parson, L.M., Murton, B.J., and Browning, P. (Eds.),Ophiolites and Their Modern Oceanic Analogues. Geol. Soc. Spec. Publ.London, 60:39-64.

McCabe, C, Jackson, M., and Ellwood, B.B., 1985. Magnetic anisotropy inthe Trenton Limestone: results of a new technique, anisotropy of anhys-teretic susceptibility. Geophys. Res. Lett., 12:333-336.

Nagata, T., 1961. Rock Magnetism: Tokyo (Maruzen).Nicolas, A., Reuber, I., and Benn, K., 1988. A new magma chamber model

based on structural studies in the Oman ophiolite. Tectonophysics,151:87-105.

Nye, J.F., 1985. Physical Properties of Crystals: New York (Oxford Univ.Press).

Osborn, J.A., 1945. Demagnetizing factors of the general ellipsoid. Phys.Rev., 67:351-357.

Pfleiderer, S., and Halls, H.C., 1990. Magnetic susceptibility anisotropy ofrocks saturated with ferrofluid: a new method to study pore fabric? Phys.Earth Planet. Inter., 65:158-164.

Richter, C, and van der Pluijm, B.A., 1994. Separation of paramagnetic andferrimagnetic susceptibilities using low temperature magnetic suscepti-bilities and comparison with high field methods. Phys. Earth Planet.Inter., 82:113-123.

Richter, C, van der Pluijm, B.A., and Housen, B.A., 1993. The quantifica-tion of crystallographic preferred orientation using magnetic anisotropy.J. Struct. Geol., 15:113-116.

Rochette, P., and Fillion, G., 1988. Identification of multicomponentanisotropies in rocks using various field and temperature values in acryogenic magnetometer. Phys. Earth Planet. Inter., 51:379-386.

Schmidt, P.W., 1993. Palaeomagnetic cleaning strategies. Phys. EarthPlanet. Inter., 76:169-17'8.

Schultz-Krutisch, T., and Heller, F., 1985. Measurement of magnetic suscep-tibility in Buntsandstein deposits from Southern Germany. J. Geophys.,57:51-58.

Smith, G.M., and Banerjee, S.K., 1985. Magnetic properties of plutonicrocks from the central North Atlantic Ocean. In Bougault, H., Cande,S.C., et al., Init. Repts. DSDP, 82: Washington (U.S. Govt. PrintingOffice), 377-383.

Stephenson, A., 1994. Distribution anisotropy: two simple models for mag-netic lineation and foliation. Phys. Earth Planet. Inter., 82:49-54.

Tarling, D.H., and Hrouda, F., 1993. The Magnetic Anisotropy of Rocks:London (Chapman and Hall).

Uyeda, S., Fuller, M.D., Belshé, J.C., and Girdler, R.W., 1963. Anisotropy ofmagnetic susceptibility of rocks and minerals. J. Geophys. Res., 68:279-291.

Wagner, J.-J., Hedley, I.G., Steen, D., Tinkler, C , and Vuagnat, M., 1981.Magnetic anisotropy and fabric of some progressively deformed ophi-olitic gabbros. /. Geophys. Res., 86:307-315.

Wolff, J.A., Ellwood, B.B., and Sachs, S.D., 1989. Anisotropy of magneticsusceptibility in welded tuffs: application to a welded-tuff dyke in theTertiary Trans-Pecos Texas volcanic province, USA. Bull. Volcanol.,51:299-310.

Date of initial receipt: 1 August 1994Date of acceptance: 12 December 1994Ms 147SR-025

403