Paleomagnetic measurements for Site 1271 were possible only on a small number of samples as a consequence of the low core recovery in both Holes 1271A (recovery = 12.9%) and 1271B (recovery = 15.3%). Remanence data were obtained from a total of 15 discrete samples and 47 archive-half pieces of altered peridotites (Tables T8, T9).
Remanence measurements were made for archive-half pieces where the vertical orientation was unambiguous and the piece length was greater than ~8 cm. Data were obtained from the centers of core pieces, treating pieces <8 cm and intervals within 4 cm of piece ends as voids. This minimum length is still smaller than the overall width of the magnetometer response functions (~8 cm width at half height), and so some artifacts are inevitable. In addition, where only a small number of suitable pieces were present in a section, these pieces were measured alone to minimize any interfering signal from adjacent pieces with differing magnetization orientations. The archive-half data are thus far from ideal, but these data nonetheless provide some indication of the stable remanence directions and intensities at the site.
The overall low recovery from Holes 1271A and 1271B makes identification of downhole trends in the archive-half data difficult. For example, Hole 1271A yielded oriented pieces of sufficient length for continuous measurements only from Cores 209-1271A-1R and 4R. Natural remanent magnetization (NRM) intensities for both Holes 1271A and 1271B were generally between 1 and 10 A/m and were measurable in the 2G magnetometer at the slowest track speed. Intensities from Section 209-1271A-1R-1 were >10 A/m, resulting in residual counts (>1000) on the z-SQUID (superconducting quantum interference device) axis. Much lower intensities (3 x 10–4 to 5 x 10–4 A/m) were measured for two pieces (Sections 209-1271B-14R-1 [Piece 4] and 17R-1 [Piece 2]) of BAG (see "Igneous and Mantle Petrology" for a discussion of the occurrence and significance of this rock type in the cores).
The remanence prior to demagnetization is commonly dominated by a substantial low-stability overprint (Fig. F69B, F69D, F69E). This steep, presumably drilling-related overprint was removed at peak demagnetizing fields of 20–30 mT. In a small number of core pieces, notably in the gabbro-impregnated peridotites from Cores 209-1271B-10R and 12R (Fig. F70D, F70E), the low-stability overprint constitutes a relatively minor portion of the remanence. After removal of this low-stability component, a variable percentage of remanence (1.5%–30% of the NRM) remained. Characteristic remanent magnetization (ChRM) directions were calculated by principal component analysis (Kirschvink, 1980) from the archive-half data over the demagnetization interval of 20 to 40–50 mT without anchoring the vectors to the origin. On average, the ChRM represents ~8% of the NRM (Table T9).
Nine samples from Hole 1271A and six samples from Hole 1271B were subjected to stepwise alternating-field (AF) demagnetization (Table T8). Five of the samples from Hole 1271A were quarter-round samples (25–50 cm3). These were demagnetized using the in-line demagnetizer and measured in a single position with the 2G magnetometer. The remaining samples (cubes with a nominal volume of ~9 cm3) were demagnetized using the off-line DTech demagnetizer and measured in three positions at each step (see "Paleomagnetism" in the "Explanatory Notes" chapter).
The NRM intensities for these discrete samples (3.4–19.5 A/m) (Table T8) corroborate the values obtained from the continuous measurements. In contrast to the pass-through measurements, complete removal of the low-stability overprint was generally achieved at AF peak fields of ~15 mT (Figs. F69A, F69C, F69D, F70A–F70C). Linear trends directed toward the origin were obtained from 20 to 50 mT, and a characteristic remanent direction was calculated by principal component analysis (Kirschvink, 1980). As observed from previous sites, low median destructive fields (MDFs) ranging 3–12 mT (Fig. F71) are the consequence of the low-field isothermal remanence induced during drilling. The highest MDF value occurs in a troctolite (Sample 209-1271B-12R-1, 131–133 cm).
Comparison of stable remanence directions obtained from discrete samples and archive halves from the same interval shows generally good agreement. The low-stability overprint is more pronounced in the archive-half data (Fig. F70), but a similar final direction (average difference = 17° ± 7.4° for 15 data pairs) is commonly isolated in both types of data. A small number of intervals show angular discrepancies near 30°. Figure F69C–F69E illustrates a progression from a shallow ChRM direction in a standard-sized (~9 cm3) discrete sample to a somewhat steeper direction in a larger quarter-round sample and the steepest direction in the corresponding archive-half data. These data might be interpreted as reflecting a more pronounced drilling-related overprint in the exterior of the core. However, we find no systematic bias toward steeper inclinations in the archive halves relative to the discrete samples for the overall data set from Site 1271. The mean inclination discrepancy for the 15 data pairs is 5° ± 15°, with the discrete samples sometimes having the steeper inclinations (cf. Tables T8, T9). Despite these differences, the overall correspondence between the discrete and archive-half data is good.
Both archive halves and discrete samples yield exclusively positive inclinations suggestive of normal polarity (Fig. F72). A single archive half (Section 209-1271B-12R-1 [Piece 8]) with a negative inclination was later shown to be inverted prior to labeling (Table T9). The mean inclination for discrete samples from Site 1271 is 25.3° (+10.6°/–13.2°; 
= 10.3; N = 15; using the inclination-only method of McFadden and Reid, 1982). The mean inclination for archive halves is 29.0° (+3.4°/–4.6°; = 25; N = 46) and is not significantly different from the average discrete sample inclination. Neither value for mean inclination is significantly different from the reference geomagnetic dipole inclination (28°), and therefore no block rotation is required since the magnetization of the rocks was acquired.
Anisotropy of the magnetic susceptibility (AMS) was measured on all 10 cube samples. The degree of anisotropy (P = maximum/minimum eigenvalues) ranges 1.06–1.33 (Table T10). All samples are statistically triaxial (i.e., all three eigenvalues are distinct), but most have magnetic fabrics that tend toward oblate shapes. Although there is significant scatter, the minimum eigenvectors for these oblate fabrics include a weak cluster in the northeast and southwest quadrants that may indicate a preferred subvertical foliation oriented northwest–southeast (Fig. F73). Three samples have more pronounced lineations (ratio of maximum/intermediate eigenvalues is larger than the ratio of intermediate/minimum eigenvalues). The AMS fabrics reflect the preferred orientation or distribution of elongate magnetite grains or grain clusters (possibly along serpentinite veins) and thus may provide an estimate of the dominant orientation of magnetite-bearing veins.
