INTRODUCTION
Many of our ideas on where mantle plumes originate, how they interact with the convecting mantle, and how plates have moved in the past rest on interpretations of the Hawaiian-Emperor hotspot track. One reason this volcanic lineament has attained this conceptual stature lies in its prominent bend at 43 Ma. The bend, which separates the westward trending Hawaiian islands from the northward trending Emperor seamounts (Fig. 1), has no equal among the Earth's hotspot tracks. It is the clearest physical manifestation of a change in plate motion in a fixed hotspot reference frame. Because the bend is so distinct, it can be used to estimate plume diameters and to place bounds on the velocity and viscosity of the convecting upper mantle that may deflect plumes (Duncan and Richards, 1991). However, shortly after hotspots were used as a frame of reference (Morgan, 1971), apparent discrepancies involving the Hawaiian-Emperor track arose (Molnar and Atwater, 1973). Attempts to model past plate motions failed to predict the bend; instead, a more westerly track was derived (Solomon et al., 1977). Tests of the fixed hotspot hypothesis suggested large relative motions between Hawaii and hotspots in the Atlantic and Indian Ocean basins (Molnar and Atwater, 1973; Molnar and Stock, 1987), but uncertainties in the plate circuits employed in these tests limited their resolving power (Acton and Gordon, 1994).Recently, several works have readdressed these questions. Norton (1995) suggested that the Hawaiian-Emperor bend records the time when the hotspot became fixed in the mantle. Prior to 43 Ma, according to Norton (1995), the hotspot moved southward, creating the Emperor seamount chain. This work is difficult to assess because of the lack of formal error analyses, but the interpretation reiterates findings of updated plate circuit studies that consider rotation pole errors (Cande et al, 1995). In addition, there is no obvious change in spreading rate at 43 Ma in the well studied marine magnetic anomaly record of the North Pacific (Atwater, 1989). Many feel the lack of such a response by neighboring plates to a change in Pacific plate motion as large as that indicated by the Hawaiian-Emperor bend is reason enough to question hotspot fixity. New modeling efforts, utilizing a viscosity structure based on geoid constraints, mantle flow fields consistent with tomographic data, and plate motion estimates also predict motion of hotspot groups (Steinberger and O'Connell, 1997). For the Emperor trend, the predicted motion is 10-15 mm/yr (Steinberger, 1996) (Fig. 2).
Whereas these recent studies have revitalized discussions regarding hotspot fixity (see also Christensen, 1998; Wessel and Kroenke, 1998), they face some fundamental data limitations. Fortunately, the hypothesis of hotspot motion can be tested independently using paleomagnetic data. If the hotspot has remained fixed (with respect to Earth's spin axis), the paleolatitudes of extinct volcanic edifices comprising the Emperor chain should match the present-day latitude of Hawaii but these tests are difficult in practice. For example, paleomagnetic data from some deep-sea sediments show a bias caused by compaction-induced flattening (Tarduno, 1990). Such problems can be avoided through the study of drill cores from well-dated lava flows. Until recently, however, only a few sites had sufficient depth penetration. This situation has improved with the latest Pacific drilling. Data from Ocean Drilling Program (ODP) Legs 143 and 144 indicate significant motions between hotspot groups in the Atlantic and Pacific Oceans during the mid-Cretaceous (128-95 Ma) (Tarduno and Gee, 1995). The motion is rapid, at speeds typical of lithospheric plates (30 mm/yr).
These findings indicate an older episode of hotspot motion and, coupled with the inferences based on relative plate motions, suggest that Hawaiian hotspot motion is a viable hypothesis that should be tested further. New data obtained from the Emperor chain drilled during Leg 145 have allowed a preliminary test. Below we summarize these analyses (Tarduno and Cottrell, 1997), as they provide support for the hypothesis and have guided our proposed sampling plan. In addition, we outline how sites chosen to address the question of hotspot fixity can provide crucial data required for understanding characteristics of the past geomagnetic field and for determining the compositional variability of volcanic products from the Hawaiian hotspot.