The nature of geomagnetic field behavior during polarity transitions is one of the most hotly debated issues in modern geophysics. Dynamo action in the earth's fluid outer core is known to be responsible for generating the earth's main magnetic field, however, the mechanism for polarity transition is still unknown. Polarity transitions are a fundamentally important property of the geomagnetic field, but because they have not taken place in historic time, it is necessary to turn to the paleomagnetic record to understand the mechanism by which they occur.
Ever since the first detailed records of transitional geomagnetic field directions were obtained [e.g. van Zijl et al., 1962], paleomagnetists have sought to understand the behavior of the field during polarity transitions. Early workers concluded that the transitional field was dominantly dipolar, based on the observation that virtual geomagnetic poles (VGPs) from polarity transitions recorded at different sites all tended to lie along the same path [ Creer and Ispir, 1970; Steinhauser and Vincenz, 1973]. A direct test of the dipolar hypothesis was made by Hillhouse and Cox [1976] who compared their record of the Matuyama-Brunhes transition (780 kyr) at Lake Tecopa, California, with a record of the same reversal from the Boso Peninsula, Japan [ Niitsuma, 1971]. The VGP paths determined from California and Japan did not overlap, and it was concluded that the geomagnetic field was not dominantly dipolar, but complex and predominantly non-dipolar, during the Matuyama-Brunhes polarity transition. As more records of a single reversal were obtained, it became evident that VGP paths were located close to the longitude of the site or antipodal to it, suggesting a transitional field that was roughly symmetrical about the earth's rotation axis [e.g. Hoffman, 1977; Hoffman and Fuller, 1978; Fuller et al., 1979]. The common-site-longitude model was readily testable and many additional transitional records were obtained. This hypothesis adequately accounted for the field behavior observed in many records, however, a growing number of records provided evidence that the transitional fields were complex with significant non-axisymmetric components [e.g. Herrero-Bervera and Theyer, 1986; Valet et al., 1988a; Clement, 1991; Clement and Kent, 1991]. Tric et al. [1991] compiled a large data set of records of different polarity transitions and short polarity events from sedimentary sequences that span the last 10 myr. They noted a visually compelling longitudinal confinement of transitional VGP paths of these records along a meridian over the Americas, and to a lesser extent, antipodal to it along a meridian over western Australia and east Asia. The records of Clement [1991] and Clement and Kent [1991] also suggest a confinement of VGP paths over the same pair of longitudes. This observation led Tric et al. [1991] and Laj et al. [1991] to suggest that dipolar field configurations may have dominated transitions over the last 10 myr.
It therefore appears, as pointed out by Bogue [1991], that we have gone full circle, with the recent revival of dipolar transitional field hypotheses. The purpose of this paper is to selectively review developments during the last quadrennium concerning our knowledge of the behavior of the geomagnetic field during polarity transitions and excursions. An excellent description of many of the concepts discussed below is given by Merrill and McFadden [1990] and by Jacobs [1994]. Other recent reviews include those by Clement and Constable [1991], Bogue and Merrill [1992], Gubbins [1994], and Jacobs [1994]. The reviews by Clement and Constable [1991] and Bogue and Merrill [1992] predate many of the developments outlined here. The review by Gubbins [1994] deals mainly with connections between polarity reversals, historical records of secular variation, developments in dynamo theory and possible effects of core-mantle interactions on the geodynamo. Jacobs [1994] gives a detailed and up-to-date account of transitional field behavior, although many of the developments outlined here occurred subsequent to the completion of his book. The emphasis of the present paper is on the paleomagnetic record of polarity transitions and the vigorous debate that has surrounded recent analyses of transitional field behavior. Geomagnetic excursions are discussed insofar as they relate to the discussion concerning transitional field behavior.