PALEOMAGNETISM

Paleomagnetic measurements were limited by poor recovery (<15% in Holes 1270A, 1270C, and 1270D) and the small lengths of most pieces. For example, only a single oriented piece could be measured in Hole 1270A. The higher recovery (37.4%) in Hole 1270B allowed sufficient archive-half measurements and discrete sampling to provide a more robust statistical characterization of the remanence. Therefore, much of the discussion will focus on results from the oxide gabbros cored in Hole 1270B (10 discrete samples and 73 core pieces measured) (Tables T9, T10). Holes 1270C and 1270D were drilled some 200 m east of Hole 1270B. These two holes are located within ~30 m of each other, and similar peridotite samples were recovered at both locations. The combined magnetic results from these two holes (10 discrete samples and 34 core pieces measured) provide a useful comparison to the data from Hole 1270B.

Susceptibility data were acquired with the MST for all suitable core samples. The serpentinized peridotites from Holes 1270C and 1270D have susceptibility values as high as 4 x 10^–2 SI. However, the discontinuous nature of these data makes it difficult to recognize any downhole trends. For the oxide-rich gabbro samples from Hole 1270B, the whole-core susceptibility data were frequently >0.1 SI, resulting in clipping of the data from the Bartington susceptibility meter. In order to identify and correct these intervals of clipped data, individual archive-half pieces were remeasured (see "Igneous and Mantle Petrology"). The smaller volume of these half-core pieces allowed most of the overrange values to be identified, and the corrected susceptibility values yielded a semiquantitative estimate of the magnetite content.

As might be expected from their high magnetite content (up to 12 vol%), the natural remanent magnetization (NRM) of Hole 1270 B archive halves was commonly too high to reliably measure even at the slowest possible track speeds (1 cm/s) in the LongCore program. Stepwise alternating-field (AF) demagnetization reveals that the initial remanence of the oxide gabbros is dominated by a steep, downward-directed, low-stability magnetization component that was likely acquired during drilling (Fig. F110). As a result of this significant low-stability overprint, stepwise demagnetization often yields vectors that lie along a great circle path between a steep (drilling induced) component and near-horizontal inclinations. Despite the substantial low-stability overprint, the remanence directions after the highest demagnetization steps (>40 mT) generally cluster at shallow negative inclinations (e.g., Fig. F110A). The remanent intensity of these high-field steps typically represents only a few percent of the NRM intensity (Table T10), and so the archive-half data alone do not provide a compelling case that a geologically significant magnetization direction has been measured. However, based on the demagnetization behavior of discrete samples, we suggest that these shallow remanence directions do, in fact, approximate the characteristic remanent magnetization (ChRM) direction.

A low-stability drilling remanence is also identifiable in archive-half data from Holes 1270C and 1270D. However, this steep component represents a much smaller fraction of the NRM (Table T10). For the peridotite samples from these two holes, stable remanence directions are commonly isolated at AF treatments above 25–30 mT. As discussed below, the remanent declinations derived from selected archive-half pieces are likely to provide a close approximation of the ChRM declination even though the inclination derived from the continuous measurements may be biased.

Hole 1270B

A total of 10 discrete samples from Hole 1270B were subjected to stepwise AF demagnetization. The NRM intensity of these samples ranges from 2.9 to nearly 50 A/m (Table T9). As observed for the archive halves, a substantial portion of this initial magnetization may be attributed to the drilling process. All discrete samples exhibit a low-stability magnetization component with steep positive inclinations (Fig. F111). This low-stability component is removed by AF treatment levels of ~15–25 mT, allowing isolation of a stable remanence direction with shallow (typically negative) inclinations in some cases (Fig. F111A). The high-stability magnetization component represents as much as 17% to as little as 0.8% of the NRM, resulting in a continuum of demagnetization behavior, from clear isolation of the ChRM to samples that apparently only record a steep, drilling-induced magnetization (Fig. F111). For the latter samples, however, a small negative inclination component may also be isolated at treatments >25 mT (Fig. F111). Indeed, the high-coercivity part of the remanence (>15–25 mT) is relatively consistent both in magnitude (0.1–1 A/m) and direction, despite the more than one order of magnitude variation in the NRM intensity.

These observations are most easily explained by the presence of two populations of remanence-carrying grains with different coercivities. Although we presently have no compositional information on the iron-titanium oxides in the Hole 1270B oxide gabbros, relatively low-Ti titanomagnetites (mean ulvöspinel content = 0.15) have been documented in similar rocks recovered from Hole 735B (Natland et al., 1991). Relatively fine grained (titano)magnetite would carry the higher-coercivity ChRM, whereas a population of coarser, lower-coercivity (titano)magnetite grains would preferentially acquire a remanence during the drilling process. The presence of two relatively discrete grain size populations is supported by the coercivity spectra obtained during demagnetization (Fig. F112). Data for all samples show curved trends, which could be interpreted as a combination of two linear trends with distinctly different slopes on a logarithmic plot of remanence decay during AF demagnetization, suggesting the presence of two discrete grain size populations. The volumetric importance of coarse-grained (titano)magnetite is also supported by the low median destructive field (MDF; the alternating field required to reduce the vector difference sum to 50% of its initial value). MDF values of ~2–5 mT (Table T9) are consistent with the presence of coarse-grained multidomain magnetite. Abundant coarse-grained magnetite is found in Hole 1270B gabbros (see "Igneous and Mantle Petrology"). In the absence of an isothermal remanence from drilling, these coarse grains would not contribute significantly to the NRM (e.g., Dunlop and ödzemir, 1997).

Holes 1270C and 1270D

The NRM intensity of peridotite samples from Holes 1270C and 1270D ranges from 0.6 to 4.3 A/m (Table T9). In contrast to the oxide-rich gabbros from Hole 1270B, peridotites recovered from Holes 1270C and 1270D have a much smaller drilling-related overprint. This steep, low-stability component is removed by demagnetization at ~6–12 mT (Fig. F113). The characteristic remanence generally accounts for >50% of the total NRM and decays linearly to the origin at fields of 50–80 mT. In contrast to Hole 1270B, the uniform slopes of the remanence decay in the log scale upon AF demagnetization (Fig. F113) suggest a unimodal (normal) distribution of magnetite grain sizes, where the contribution of unstable coarse-grained magnetite is negligible. As a result, observed MDF values (10–20 mT) are systematically higher than in the oxide-rich gabbros of Hole 1270B.

Despite the small number of samples, most of the paleomagnetic directions from Hole 1270B have negative inclinations. Results from discrete samples range from –35° to +5°, whereas long-core continuous measurements show a wider variance, typically from –20° to +15° (Fig. F114). The discrepancy between discrete and long-core inclinations is more obvious in Holes 1270C and 1270D, where the characteristic component represents a larger proportion of the NRM (Fig. F113; Table T9) and calculation of paleomagnetic directions has less error. For example, the demagnetization data from the archive halves often show nearly unidirectional demagnetization trends at the same depth interval where off-line treatment of discrete samples yields a distinct two-component magnetization (Fig. F114). This discrepancy may reflect heterogeneities in the archive-half magnetization (note that there is a factor of 3 difference in the NRM intensity). Alternatively, the drilling overprint may be more pronounced in the outer portion of the core, leading to a relatively larger contribution in the archive halves than in the discrete samples (taken closer to the center of the core). Given the different demagnetization behavior in the archive-half and discrete sample data, two points should be mentioned. First, because the drilling-related low-field overprint is nearly vertical, declinations are not excessively affected by a partial overlap of the drilling component. Stepwise demagnetization yields great circles aligned about nearly vertical planes (see Fig. F110). Thus, the half-core declinations generally show good agreement with discrete sample results (Fig. F115). For samples from Holes 1270C and 1270D, the two types of data are always within 20° of each other. Somewhat larger discrepancies between archive-half and discrete sample declinations are evident for Hole 1270B gabbroic samples, which have a substantial drilling-induced overprint. Where both types of data are available the paleomagnetic directions obtained from discrete samples should be regarded as more accurate (see "Paleomagnetism" in the "Site 1268" chapter).

The drilling-induced remanence precludes the use of a north-directed recent viscous component for core reorientation. In the absence of azimuthal constraints, the polarity interpretation of the directional data relies exclusively on the inclination of the remanence. In the case of Hole 1270B, inclinations are shallow and predominantly negative. Moreover, the mean inclination for 10 discrete samples is –14° (+10°/–10°; = 25) using the inclination-only technique of McFadden and Reid (1982). These data are most easily interpreted as representing a reversed polarity magnetization acquired during the Matuyama chron (older than 0.78 Ma) (Cande and Kent, 1995). On the other hand, the combined data set from Holes 1270C and 1270D yields a mean inclination of –3° (+13°/–13°; = 15; N = 10) for discrete samples and a distribution (including both discrete and archive-half data) that is slightly skewed toward positive inclinations. The polarity interpretation is therefore ambiguous. Perhaps the simplest scenario would be that all holes from Site 1270 have the same reversed polarity. Alternatively, the different lithologies recovered from Hole 1270B (oxide gabbros) and Holes 1270C and 1270D (harzburgites) and the spacing between these holes suggest another possible interpretation; namely, low-temperature (~300°C) serpentinization and remanence acquisition in the peridotites might represent a normal polarity magnetization acquired at a somewhat later time than the magnetization in the gabbros.

The mean paleomagnetic inclinations of the gabbros and the harzburgites are not significantly different. However, their values are significantly lower than the expected inclination for the time-averaged dipole field at the site (±28°). As suggested for Site 1268, rotation about a ridge-parallel, near-horizontal axis may have affected Site 1270. Because the remanence of Hole 1270B is interpreted as reversed polarity, an alternative explanation for the shallow remanent inclination would be that the drilling overprint has overlapped the characteristic magnetization. In contrast, the drilling remanence and characteristic component are well separated in discrete samples from Holes 1270C and 1270D (Fig. F113). Thus, the shallow inclinations for these holes apparently require some rotation. The uncertainty of the magnetic polarity of samples from Holes 1270C and 1270D does not affect this conclusion, but it does affect the reorientation of any structural data.

Declinations determined from both archive halves and discrete samples from Hole 1270B and Holes 1270C and 1270D are not randomly distributed in the core reference frame (Fig. F116). Although the cores are not azimuthally oriented, the dominant foliation was used to determine the splitting line for the cores. The clustering of declination data therefore indicates the general success of this procedure and the existence of a relatively uniform foliation plane. The declinations for Hole 1270B are near 200°, while the declinations from Holes 1270C and 1270D cluster near 350°. This difference indicates that either the foliations do not have a common attitude or the polarity in the two locations is different (see "Structural Geology" for a complete discussion).

Anisotropy of magnetic susceptibility (AMS) was routinely measured on all discrete samples (Table T11). The degree of anisotropy (P = maximum/minimum eigenvalues) ranges from 1.03 to 1.19. The majority of samples have an oblate magnetic fabric. The bulk susceptibility values for all samples are sufficiently high (>10^–3 SI) that the contribution from paramagnetic silicate grains can be neglected (Rochette et al., 1992). The AMS signal reflects either the shape anisotropy of grains (or grain clusters) or the anisotropic distribution of magnetite. Although the precise relationship of the magnetic anisotropy to the silicate fabric in the Site 1270 samples is presently not known, the poles to magnetic foliation (i.e., the minimum eigenvector of the susceptibility tensor) are broadly consistent with the mesoscopic foliations (Fig. F117).