CONCLUSIONS

In this initial study of the Japan Trench volcaniclastic sediments, we have attempted to ascertain the potential of the Pliocene-Pleistocene tephras to provide a tephrochronological framework for the Japan area and a record of arc evolution. This approach reflects global developments (e.g., see Lowe, 1996; Hunt, 1999) that are using high-precision tephrostratigrapical studies to refine our knowledge of the chronology of environmental change and the linkages between marine and terrestrial geosystems. We draw the following conclusions in relation to the Japan Trench tephras.

1. Although there is convincing evidence that hydration (phase 1 alteration) of glass shards increases with age in the Pliocene-Pleistocene, there is no indication that element leaching (phase 2 alteration) has occurred. Provided that inter-environmental differences in hydration rates are considered, the potential for correlation with tephras elsewhere is not hindered by pervasive diagenesis during Pliocene-Pleistocene times.
2. There are no differences in the geochemical homogeneity of glass shard populations in pods and layers. The use of tephra pods as potential stratigraphical markers is therefore possible.
3. Despite high-precision analysis of glass shards, it is in some instances not possible to distinguish geochemical differences between tephras (eruptions). This presents problems in the marine record where the low number of coring sites and the inability to perform "hand-over-hand" lateral correlations between closely spaced sites prevents unambiguous tephra correlation. Therefore, correlation of tephra layers cannot be achieved by geochemical fingerprinting alone but requires systematic high-resolution stratigraphical and geochemical analysis of all visible and crypto- tephra layers and/or independent geochronological data in parallel with the tephrostratigraphy.
4. The relatively high recovery of Pleistocene sediments can permit, for the first time, a near-continuous temporal record downwind from centers of explosive volcanism on Honshu in particular and the Japan arc in general. An age model that accords with generally accepted bio- and magnetostratigraphy is, however, required before this can be achieved. We maintain that tephrochronology should not be a critical input to this model, unless all tephras in a particular record have been analyzed.
5. The general homogeneity of the pods and layers and the relatively nonpervasive alteration suggests that the volcaniclastic record of the Japan Trench is suitable for further geochemical investigations of arc evolution.