CONCLUSIONS

1. The characteristic remanent magnetization directions from u-channel measurements and single samples at Site 1020 are well defined before and after the magnetic reversals. The determination of ChRMs for transitional samples is more difficult because of their low intensities and the badly defined primary magnetization directions. The negative inclinations at the northern Site 1020 are slightly shallower than the value expected at this latitude for an axial geocentric dipole field. This suggests an unremoved coring-induced overprint whose coercivity spectrum overlaps with the coercivities of the ChRM. The Brunhes-Matuyama transition at the southern Site 1014 shows shallow negative and steep positive inclinations.
2. The VGP paths for the u-channel measurements and the discrete samples from the upper Jaramillo transition at Site 1020 are roughly similar. The VGPs from the u-channels track across the Pacific Ocean, whereas the only transitional VGP from the discrete samples lies over the Gulf of Mexico. The B-M reversal at Site 1020 shows differences between the u-channel measurements and the discrete sample results, so the interpretation of the transitional VGPs must be handled with caution. The effect of the drilling-induced overprint on the VGPs at Site 1020 can be seen in both records (Fig. 4, Fig. 6). In future studies, thermal demagnetization should be tried if AF treatment fails to remove the hard drilling remanence.
3. Variations in grain size and composition of magnetic particles along the core are small at Site 1020.
4. The Brunhes-Matuyama transition at Site 1014 in the Tanner Basin demonstrates convincingly the persistent nature of the drilling-induced overprint. In that case the VGPs do not reach the geographic North or South Pole. As a result of the two magnetic minerals and the NRM intensity drop in the sediments of Site 1014, the directional record of these sediments should only be regarded as an illustration of a persistent coring-induced magnetization.
5. Because of the hard coring-induced remanence components, we do not consider any of the transitional records from this study suitable for extracting details about the geomagnetic field.
6. Model calculations for the VGP paths of the upper Jaramillo transition show that incompletely removed radial and vertical overprints can strongly affect the track of the VGPs. The antipodality of normal and inverse directions can be improved by subtracting a radial and vertical component of magnetization. Previously measured records from ODP cores must be investigated for the possible presence of coring-induced overprints.