It was our objective to determine the paleoinclination of the Site 1213 sills. We found many samples with a consistent shallow negative inclination in the range –0.8° to –26°, but usually between –5° and –15°. These inclinations are interpreted as the initial, characteristic remanence magnetization recorded in the sills. We do not think the positive-inclination samples are representative of the true magnetization direction for several reasons. Most samples are overprinted by a steep, downward-pointing magnetization imparted by the drill string (e.g., Acton et al., 2002). In most of the positive-inclination samples, the overprint and final characteristic magnetization direction appear to overlap, with a gradual shift from the former to the latter during progressive demagnetization. This observation suggests that the overprint may not be completely removed. Furthermore, the positive inclinations show more scatter than the negative (Fig. F5), which is consistent with variable removal of the drill string overprint. In addition, the positive inclinations are mixed with negative inclinations in Unit 2 (Fig. F5). This observation implies that positive-inclination values do not represent a separate unit with a different magnetization direction but are instead a spurious direction. In Unit 1, the positive inclinations are grouped, and we could interpret them as a separate unit, but the simplest explanation is that these inclinations are samples in which the downward, positive drill string overprint has not been completely removed.
Mean inclinations in the three sills are not statistically distinct. The similarity in the paleoinclination of the three units probably means that the units erupted during a short time interval. Averaging all 26 samples with reliable negative inclination values gives a mean inclination of –9.3°. Although the tight clustering of inclination values results in a standard deviation of only 4.3°, the low scatter is not indicative of the true accuracy of the paleoinclination because inclination variation caused by secular variation has not been averaged out. The Cox and Gordon (1984) method uses a model of secular variation to propagate the uncertainty owing to secular variation into uncertainty estimates. The model standard error for secular variation at Hole 1213 is 9.1° in colatitude (Cox and Gordon, 1984). When combined with an estimate of 2° for the uncertainty in the vertical orientation of the borehole and the observed scatter in inclination values (Cox and Gordon, 1984), this method gives 95% confidence limits spanning 69° for paleoinclination (–41.8° to 27.5°) and 38.6° for paleolatitude (–24.0° to 14.6°). This large uncertainty stems from the fact that all of the measurements amount to a single spot reading of the geomagnetic field and thus paleosecular variation is poorly averaged.
Demagnetization and hysteresis experiments suggest that the Hole 1213B basalts contain low-coercivity magnetite grains that easily acquire an overprint, probably from the drill string, resulting in high NRM values. Interestingly, our results are similar to those from sills cored at Deep Sea Drilling Project (DSDP) Site 462 (Steiner, 1981). This suggests that something about sills, perhaps a long cooling history that results in larger magnetic grain sizes, may cause unreliable magnetite behavior.
Because of the small number of units and low latitude, we cannot be certain of the polarity and true inclination. With a negative mean inclination, the polarity is reversed if the lithosphere was formed north of equator or normal if it was formed south of the equator. The explanation most consistent with other observations is a reversed polarity, formed north of the equator. Reversed polarity is consistent with the magnetic model of the Southern High, which gave a negative magnetization (Sager and Han, 1993). Furthermore, assuming reversed polarity, the colatitude arc is in better agreement with Late Jurassic and Early Cretaceous paleomagnetic poles determined from anomaly skewness (Larson and Sager, 1992), Jurassic and Early Cretaceous sediment data from the western Pacific (Steiner and Wallick, 1992), and other Pacific basalt core data (Fig. F6). Although the large uncertainty for Hole 1213B basalts does not allow us to rule out a normal polarity, the reversed polarity is a more consistent with other data.