RESULTS

Site mean values of the mean susceptibility,

(₁ + ₂ + ₃)/3,

and NRM intensity for both pillowed and massive units are given in Table T1. Although the mean susceptibility of individual samples varies from 0.003 to 0.06 SI units, most samples have values in a much narrower range (i.e., 0.01-0.03 SI units). In contrast, NRM intensities vary over two orders of magnitude (Fig. F2). There is a tendency for the more massive units to have somewhat higher susceptibilities and NRM intensities, but differences with pillowed units are relatively minor and do not appear to be pervasive. For example, samples from the massive units at Site 1187 have lower NRM intensities than the pillow samples (Table T1). Site to site comparison, however, is limited by the uneven sample distribution between pillowed and massive units at the individual sites. For example, 90% of the samples at Site 1187 are from pillowed units. A majority of samples have high Koenigsberger ratios (Q), reflecting the well-known dominance of remanent magnetization in oceanic basalts (Fig. F2). Basaltic material obtained at Site 1185 was divided into upper and lower groups based on significant geochemical differences (Shipboard Scientific Party, 2001). These differences are also reflected in magnetic properties, particularly the mean susceptibility values, which are notably larger for the lower group.

Magnetic fabric parameters (lineation, L, foliation, F, and shape factor, T) show considerable variation within individual sites. However, a majority of samples from all sites have low lineation (1.0 < L < 1.01) and foliation (1.0 < F < 1.01) values, indicating that most samples have low anisotropy with neither strongly oblate nor strongly prolate character (Fig. F3A). Figure F3B shows that for those samples that exhibit larger lineation and foliation values there is a minor tendency for the more massive units to have higher lineation values and therefore to be more prolate. Of the massive unit samples, 57% have prolate and 43% have oblate shapes, whereas pillow samples are almost equally distributed between prolate (49.4%) and oblate (50.6%) shapes.

The degree of anisotropy (A),

100 x [1 - (₂/2₁) - (₃/2₁)]

(Canon-Tapia, 1992, 1994), of Leg 192 basalts also shows considerable variability, but a majority of samples have low values (A < 2). Although approximately one-third of the samples (92/277) have values of A > 2, only 13 (~5%) have values of A > 5. The massive units are characterized by slightly larger values of A, but several pillowed units have values of A > 5, indicating that differences between massive and pillowed units are again relatively minor. Large values (i.e., A > 5) have been linked to the anisotropy produced during rapid cooling in high-stress environments such as those found at the boundaries of flows or pillows (Canon-Tapia and Pinkerton, 2000). As such, large values of A may be used to define flow boundaries. Some samples with large A values are close to such boundaries identified while drilling during Leg 192, but many are not.

The directions of maximum, intermediate, and minimum susceptibility (i.e., ₁, ₂, and ₃) axes were examined in order to detect the presence of any preferred orientation of key fabric features such as lineation and foliation. Directions in the vertical (i.e., inclination) and horizontal (i.e., azimuth) planes were examined separately because the cores were not azimuthally oriented when drilled and their absolute orientation in the horizontal plane is therefore more uncertain than that in the vertical plane.

Inclinations

Figure F4 shows histograms of the inclination of the ₁, ₂, and ₃ axes for pillowed and massive units for all sites. Both pillowed (Fig. F4A) and massive (Fig. F4B) units show a tendency to shallow inclination (<20°) for the ₁ axis. Shallow ₁ inclinations are observed in both oblate and prolate samples and in those with small and large degrees of anisotropy, A. Similarly, in many cases the ₂ axis is also shallow. In the case of pillowed units, the ₃ axis appears to be almost randomly oriented in the vertical plane with just a minor suggestion of a bimodal distribution of shallow (<15°) and steep (~70°) inclinations. In the case of the massive units, the ₃ axis shows a definite bias toward shallow angles. In addition, several massive samples have steep ₁ axes—a feature less prevalent in the pillowed samples.

The dip of the ₁-₂ plane was calculated and is shown in histogram form for both pillowed (Fig. F4C) and massive (Fig. F4D) samples. Clearly, the more massive units are dominated by steeply dipping ₁-₂ planes with almost one-half (50/111) having dips >70°. Samples from the massive units with steeply dipping ₁-₂ planes and shallowly inclined ₃ axes (Fig. F4B, F4D) have a mostly prolate fabric. Thus, for many of the massive units the ₁-₃ or the ₂-₃ plane is nearly horizontal and is thought to represent the flow plane. In contrast, dips of the ₁-₂ plane for pillowed samples show a bimodal distribution with both shallow (<30°) and steep (>60°) dips (Fig. F4C). For many of the pillowed samples the ₁-₂ plane is subhorizontal, which together with the dominance of shallow ₁ axis (Fig. F4A) suggests that for many pillowed samples it is the ₁-₂ plane that is the flow plane.

Azimuths

Rotary drilling during Leg 192 produced cored pieces of basalt that were not azimuthally oriented. To obtain information regarding any preferred azimuthal direction that may be associated with their flow, the basalts must first be oriented with respect to a fixed common direction. This was done using the ChRM to define magnetic north at the time of basalt eruption. The ChRM direction for most samples has been well defined using standard alternating-field and thermal demagnetization techniques (Riisager et al., 2003). However, the directions for individual samples also reflect the influence of secular variation and, as such, do not provide a single fixed direction but rather give an approximate direction to magnetic north defined by an axial dipole. The range of magnetic declination is thought to be small (<15°) and is not therefore further accounted for in the directional analysis. However some of the variation in flow directions described below may be attributable to secular variation.

Preferred azimuths were initially determined for both pillowed and more massive units using the ₁ direction. For samples with shallowly inclined (<15°) ₁ directions, the preferred direction was selected as the ₁ azimuth. For samples with steep ₁ inclinations—dominantly those associated with samples from the massive units—the ₂ direction was chosen as the preferred azimuth. In those cases where the ₁ and ₂ axes have intermediate inclinations but lie in an almost vertical plane (i.e., where the ₁ - ₂ azimuthal difference is 180° ± 20°), the strike of the vertical plane (i.e., a line perpendicular to the ₃ azimuth) was chosen as the preferred azimuth. In samples where the ₁ axis was neither shallow (<15°) nor steep (>75°) and the ₁-₂ plane was not near vertical, no preferred azimuth was assigned. Approximately 50% of the samples from each site met these criteria and were used to determine a preferred azimuth. Because the ₁ (or ₂) axes of these samples are all shallow (<15°), it is difficult to unambiguously determine which of the two possible antipodal directions is correct. Accordingly, all of the individual azimuths have been rotated into the either the southeast or the southwest quadrant and displayed as Rose diagrams (Fig. F5). The mean preferred direction for each site was determined using a simple Gaussian distribution for all azimuth data and a Fisher distribution for the inclinations and declinations of the ₁ (or ₂) axes (Table T2).

Ten samples obtained from a single 1.3-m-thick pillow basalt at Site 1187 were analyzed in detail. The remarkably uniform ChRM directions (<2° variation in inclination and ~5° variation in declination) (Fig. F6) are consistent with the jigsawlike fit of the individual pieces, indicating that the pillow is more or less intact. This consistency in the ChRM directions suggests that there has been no significant internal disruption of the pillow since it acquired its magnetization, although the entire pillow may have been tilted subsequent to magnetization. The uniformity in the paleomagnetic results also suggests that sampling pillows at points anywhere in their interiors will likely produce representative and reliable ChRM directions. The NRM intensity, mean susceptibility, and ARM intensity are relatively uniform, with minor decreases in those samples near the top and bottom reflecting changes in grain size. As shown by Figure F6, samples >15 cm from either the top or the bottom have very uniform properties. This is also true for many of the magnetic fabric parameters (e.g., lineation, foliation, and degree of anisotropy), which exhibit relatively uniform values throughout the pillow. The AMS fabric directions, on the other hand, are more variable, with substantial changes in the orientation of the ₁-₂ plane toward the pillow center. In the upper 50 cm and lower 15 cm, the ₁ axis is consistently <25° and the ₁-₂ plane has a subhorizontal (<30°) attitude. Although the ₁ axis remains shallowly inclined throughout, the dip of ₁-₂ plane changes substantially in the central portion to almost vertical (Fig. F6). In this same depth interval, between 60 and 90 cm from the top of the pillow, the dip direction (i.e., azimuth) of the plane undergoes a roughly 90° counterclockwise rotation. As shown in Figure F6, these changes in magnetic fabric directions appear to have no measurable influence on the ChRM.