A total of 52 samples from Hole 1213B were measured in this study (Table T1). Samples were obtained as 2.5-cm-diameter minicores drilled perpendicular to the split face of the rock cores. Only larger core pieces, believed to be vertically oriented, were sampled. These samples were spaced at irregular intervals in rock sections, with the objective of collecting about the same number from each sill.
Sample magnetizations were measured with a Geofyzika JR-5A spinner magnetometer. Both alternating-field (AF) and thermal demagnetization were carried out on different suites of samples to remove overprint magnetizations and to isolate the characteristic remanent magnetization. For the AF method, samples were demagnetized with a Schoenstedt Instruments AC demagnetizer model GSD-1 and for thermal treatments, samples were demagnetized with a Schoenstedt Instruments model TSD-1. All measurements were made within shielded rooms at the University of Houston, Texas (USA), paleomagnetic laboratory.
Approximately 60% of samples were treated with the AF method, usually with 2.5- to 5.0-mT steps up to 30–40 mT. Higher steps were not often used because most samples gave inconsistent directions when treated with higher fields. The rest of the samples were treated with thermal demagnetization, typically with 50°C steps in the range of 150° to 600°C, except for 12 samples that were examined in detail from 500° to 580°C with 20°C steps to look at directions above 500°C. See the "Supplementary Material" contents list for AF and thermal demagnetization data tables.
Hysteresis loops for 10 representative samples were measured using a Princeton Instruments Micromag model 2900 alternating-gradient magnetometer at the University of California, Davis (USA), paleomagnetic laboratory. Volume susceptibility was measured at Texas A&M University, College Station, Texas (USA), with a Bartington MS-2 susceptibility meter.
Characteristic magnetization directions were calculated using principal component analysis (PCA) (Kirschvink, 1980). We chose two to six (average of four) steps that appeared to have univectorial decay toward the orthogonal vector diagram origin. Sample directions were calculated using PCA with the solution not anchored to the origin when that method gave a solution apparently consistent with the observed demagnetization trends in orthogonal vector plots. If it did not, we used a PCA constrained to pass through the origin.
Individual sample magnetizations were averaged using the method of Cox and Gordon (1984). Following this method, paleoinclinations were treated as colatitudes and averaged for each lithologic unit and unit means were tested for statistical distinctness. The mean paleocolatitude was calculated and corrected for bias inherent in azimuthally unoriented data, and 95% confidence limits were estimated. In the calculation of confidence limits, the Cox and Gordon (1984) method includes an estimate of error caused by secular variation as well as off-vertical tilt of the borehole (assumed 2° or less).