In her review of sedimentary records of relative paleointensity, Tauxe [1993] briefly discussed seven relative paleointensity records of the Matuyama-Brunhes transition. Tauxe [1993] notes broad similarities between the records in that the transition zones are associated with low relative paleointensities. Valet and Meynadier [1993] then produced the first detailed relative paleointensity record that continuously spans the last 4 myr and which crosses numerous polarity reversal boundaries. The most significant feature of this record is the gradual decrease in field intensity prior to most polarity transitions and the rapid rise to high intensity after each transition. This picture has been confirmed by duplicate records, from three ocean basins, that cross the Matuyama-Brunhes transition [ Valet et al., 1994], as well as from the lower and upper Jaramillo and Matuyama-Brunhes transitions from the central North Pacific Ocean [ Verosub et al., ms]. Bogue and Paul [1993] obtained absolute paleointensity data from a sequence of Hawaiian lava flows which were erupted immediately following a geomagnetic reversal. They note that field intensities were low during the transition and unusually high in the interval immediately following the reversal, similar to the behavior observed in the Steens Mountain (Oregon) volcanic sequence [ Prévot et al., 1985]. A coherent picture therefore seems to be emerging that indicates that field intensities decay during a polarity chron, culminating in low values during the polarity transition, followed by a rebound to high intensities in the interval immediately after the transition. Nevertheless, the details of this asymmetrical saw-tooth paleointensity pattern of Valet and Meynadier [1993] will remain the subject of debate until such behavior is more widely observed.
These studies demonstrate the importance of obtaining the paleointensity which is usually the missing third component of paleomagnetic studies. Mary and Courtillot [1993] have stressed the importance of this component because it allows a full three-dimensional (3D) representation of geomagnetic polarity transition records. Mary and Courtillot [1993] present 3D polarity transition data in local cartesian coordinates (analogous to vector component or ``Zijderveld'' projections) which allows the examination of polarity transition records for systematic behavior. Lin et al. [1994] analysed four of the best known volcanic polarity transition records and one excursion record from which high quality absolute paleointensity data are available. They found a significant correlation between the virtual dipole moment and the angle between the VGP and the earth's rotation axis. This result suggests that physical models for geomagnetic field behavior must address the question of how changes in direction are linked to changes in intensity during polarity transitions.
Oppenheim et al. [1994] obtained transitional field directions from a series of lower Carboniferous lavas which yield paleointensity estimates that lie between 10 and 20% of other estimates of the lower Carboniferous field intensity. This suggests that Paleozoic geomagnetic polarity reversals were similar to more recent polarity transitions in that they involve a major decline in field intensity. Low field intensities are also associated with low latitude VGPs from the Summer Lake excursion record in Pleistocene sediments [ Roberts et al., 1994] and an excursion from volcanic rocks in Germany [ Schnepp and Hradetsky, 1994]. The directional signature of the Summer Lake excursion [ Negrini et al., 1994] is attenuated with respect to the other records of the same excursion [cf. Herrero-Bervera and Helsley, 1993]. Negrini et al. [1994] postulate that this may be due to overprinting by the subsequent high-intensity field, as has been proposed by Coe and Liddicoat [1994] for the most detailed record of the Mono Lake excursion in the Mono Lake Basin of California [ Liddicoat, 1992]. Alternatively, the directional attenuation may result from sedimentary smoothing due to the relatively low sedimentation rate at Summer Lake with respect to Pringle Falls.