CONCLUSIONS

Shipboard long-core paleomagnetic measurements made during Leg 172 identified a reproducible pattern of directional secular variation within oxygen isotope Stage 3 that could be correlated between Sites 1061, 1062, and 1063, which span a distance of more than 1200 km. Two intervals of excursional paleomagnetic directions were identified at all three sites within Stage 3 and were labeled as excursions 3 and 3. New analysis of excursions 3 and 3, based on the shipboard measurements, new U-channel paleomagnetic records of the excursions, and independently published discrete sample paleomagnetic records of the same intervals all indicate that excursion 3 is not real; its directional variability is less than originally estimated and not excursional. Excursion 3 is real and easily identified in both U-channel and discrete sample measurements. The best U-channel record of excursion 3 from Hole 1063C is almost identical to the best discrete sample paleomagnetic records. The U-channel records from Holes 1061B and 1061C are similar to one another but somewhat different from the other records (more smoothed). We attribute that difference to systematic biases in the U-channel measurement process when large-amplitude, fast directional changes occur.