We present a new palcomagnetic record of the Cobb Mountain Subchron obtained from deep-sea sediments cored at Ocean Drilling Program site 647 in the southern Labrador Sea. The details of the transitional field behavior documented by this record appear to be very similar to those recorded in a previously published record of this subehron obtained at Deep-Sea Drilling Project (DSDP) site 609 in the North Atlantic (Clement and Kent, 1987). We used a quantitative correlation technique (Martinson et al., 1982) to establish statistically the degree of similarity between these two records and thereby constrain the spatial variability in these transitional fields.The error in the alignments is reduced significantly by aligning records of virtual geomagnetic pole positions rather than directions, indicating that these records document such large scale changes in the fields that we can not distinguish them from dipolar changes, given the proximity of these two sites. These replicate records of the Cobb Mountain Subehron provide evidence that deep-sea sediments are capable of providing high resolution records of geomagnetic field behavior. A reexamination of the sequence of polarity transitions recorded at DSDP site 609 in light of these results suggests the presence of two preferred transitional field configurations. The field appears to change from one configuration to the other for several reversals and then back to the original configuration, suggesting that a geographical influence on the reversal process persists through this sequence. The variability in these reversal records provides insights into the response of the geodynamo to this geographical influence.
Whileremarkable similarities exist between these two records, there also exist obvious differences. In order to better understand the fidelity with which these sediments have recorded geomagnetic field behavior, we used a quantitative correlation technique [Martinson et al., 1982] to examine the extent to which the differences between these records could be explained by relative sedimentation rate changes between the two sites. In applying this technique we assume that the two records are the same, allowing for differences in natural noise Paper number 91JB02584.0148 -0227/92/91 lB-02584505.00and sedimentation rates as well as some differences in the actual intensities and directions. We test the degree of fidelity of these records by comparing the correlation coefficient between the records after removing the effects of relative sedimentation rate changes using the alignment technique. In aligning the records we compare the agreement between the best fitting alignment of each of the vector components. A change in sedimentation rate must affect all the vector components identically, and therefore differences in the mapping functions describing the alignments indicate that differences exist in the records which can not be explained by relative sedimentation rate changes alone. The remaining differences must result either from local factors which affected the rema...