Two alternative interpretations of the Site 907 magnetic stratigraphies were presented in the Leg 162 Initial Reports volume (Shipboard Scientific Party, 1996a). One of these interpretations (Interpretation 1 in Fig. 8) implies hiatuses in deposition at ~105 mbsf and at ~140 mbsf. The other interpretation (Interpretation 2 in Fig. 8) implies a more uniform sedimentation rate without major hiatuses. Interpretation 1 is essentially the same as the interpretation given for Hole 907A in the Leg 151 Initial Reports volume (Shipboard Scientific Party, 1995).
Shipboard biostratigraphy had insufficient resolution to distinguish between the two alternative magnetostratigraphic interpretations. Except in the top few cores, calcareous microfossils are absent. Silicoflagellates and diatoms are preserved at Site 907, particularly in the 50-100 mbsf interval, but shipboard study was too cursory to provide useful biostratigraphic constraints. Since the Leg 162 cruise, detailed shore-based studies on diatoms from Hole 907A have led to a correlation of middle Miocene to late Pliocene diatom events with the polarity time scale (Koç and Scherer, 1996). These authors modified the original magnetostratigraphic interpretation at Hole 907A in making these biomagnetostratigraphic correlations.
Neogene diatoms evolved rapidly in the high-latitude oceans, and therefore have high potential for biostratigraphy. However, diatoms of the Norwegian-Greenland Sea are endemic, and correlations established in the North Atlantic or North Pacific are not necessarily applicable to the Norwegian-Greenland Sea. No diatom zonation was established from Leg 104 (Vøring Plateau) sediments. The few diatom events documented on the Vøring Plateau cannot be unambiguously correlated with polarity chrons because of uncertainties in pre-Pleistocene magnetostratigraphic interpretations. At Hole 907A, Norwegian-Greenland Sea diatom biostratigraphy has taken a giant leap forward with the establishment by Koç and Scherer (1996) of eleven diatom "intervals" (zones) from middle Miocene to Holocene. Hole 907A now provides a reference section for diatom stratigraphy from the Norwegian-Greenland Sea, and it is important to establish the correlation of this zonal scheme with the polarity time scale. The correlation of the diatom "intervals" with the polarity time scale advocated by Koç and Scherer (1996) is based on a reinterpretation of the Hole 907A (Leg 151) magnetostratigraphy (Shipboard Scientific Party, 1995).
The endemic nature of Norwegian-Greenland Sea diatoms, the lack of existing correlations of diatoms with the polarity time scale in this region, and uncertainty in the interpretation of the Site 907 magnetostratigraphy (see Fig. 8) means that there are few constraints on possible biomagnetostratigraphic correlations. Silicoflagellates are well preserved throughout the middle Miocene-Pliocene interval at Site 907 (see Amigo, Chap. 5, this volume), and they provide a means of constraining the magnetostratigraphic interpretations. Silicoflagellates are generally less endemic than diatoms, and silicoflagellate datums recognized at Site 907 can be used to "pin" the magnetostratigraphic interpretation. Site 982, drilled during Leg 162 on the Rockall Plateau, has proved useful in this respect. This site has well-preserved silicoflagellate and calcareous nannofossil records (Amigo, Chap. 5, this volume; Shipboard Scientific Party, 1996b), allowing a correlation of critical silicoflagellate datums with nannofossil zones and hence with the polarity time scale.
In Hole 907B, the last occurrence (LO) datums of silicoflagellates Bachmannocena apiculata apiculata and Bachmannocena apiculata curvata appear together at 190.26 mbsf with the first occurrence (FO) of Bachmannocena circulus apiculata found just below, at 193.79 mbsf. In Hole 982B, the same occurrence overlap is present in the 444-454 mbsf interval, and the short-ranging species Distephanus stauracanthus occurs within the overlap interval at 449 mbsf. This unique and stratigraphically significant juxtaposition of silicoflagellate datums has been recognized at many locations; the closest geographically (to Site 907) are Deep Sea Drilling Project (DSDP) Site 408 in the North Atlantic (Bukry, 1979) and ODP Site 642 on the Vøring Plateau (Locker and Martini, 1989). For Site 408 and Hole 982B (Bukry, 1979; Shipboard Scientific Party, 1996a), the overlap of the silicoflagellate datums can be correlated with the lower part of nannofossil Zone CN5, close to the CN4/CN5 nannofossil zonal boundary (nannofossil zonation of Okada and Bukry, 1980). According to Berggren et al. (1995), the CN4/CN5 nannofossil zonal boundary lies close to the C5ABr/C5ACn polarity chron boundary.
In Hole 982B, the CN5/CN6 nannofossil boundary (= middle/late Miocene boundary) occurs at 377 mbsf, based on the LO of the nannofossil Coccolithus miopelagicus (Shipboard Scientific Party, 1996b). According to Berggren et al. (1995), this nannofossil zonal boundary correlates with C5r.2r. The FO of the silicoflagellate Bachmannocena diodon nodosa at 408 mbsf in Hole 982B is, therefore, older than C5r.2r. In Hole 907B, the FO of Bachmannocena diodon nodosa occurs at 152.40 mbsf within a reversed polarity zone (Fig. 5), which must be older than C5r.2r.
With these constraints provided by the silicoflagellate datums, we propose a correlation of polarity zones at Site 907 with polarity chrons (Fig. 3, Fig. 7) that differs from those offered in the Initial Reports volumes (Shipboard Scientific Party, 1995, 1996a). The new interpretation (Interpretation 3 in Fig. 8) is almost identical to the reinterpretation of the Hole 907A magnetic stratigraphy offered by Koç and Scherer (1996). As mentioned above, the endemic nature of diatoms from the Norwegian-Greenland Sea indicates that diatom-magnetostratigraphic correlations from outside the region may not provide useful constraints. However, one species common to the Norwegian-Greenland Sea and the North Pacific, the short-ranging diatom Denticulopsis praedimorpha, is found in the 160-163 mbsf interval in Hole 907A, correlating with C5AAn, and in the same polarity chron at Site 884 (North Pacific) (Barron and Gladenkov, 1995).
The magnetostratigraphic interpretation indicates a hiatus (disconformity) in the 100-115 mbsf interval (Fig. 3, Fig. 8) in which deposition slowed dramatically or stopped for ~2.5 m.y. There are two additional lines of evidence for this hiatus. First, a series of sawtooth-like "steps" in GRAPE density between 112 and 103 mbsf, which overprint a decreasing trend in density from 118 to 100 mbsf (Fig. 9), are associated with a change in character of the magnetic susceptibility record, which shows anomalously low values over the same depth interval. These profiles may be interpreted to suggest either (1) rapid variations in lithology and induration within this interval (which may be related to changes in sediment accumulation rates) or (2) reversals in downhole geotechnical gradients caused by erosion and/or nondeposition of sediment. Second, Koç and Scherer (1996) noted that four diatom species-Coscinodiscus norwegicus, Cymatostra biharensis, Denticulopsis hustedtii, and Ikebea tenuis-have their last occurrences at in the 113-115 mbsf interval at Hole 907A. As noted by Koç and Scherer (1996), this juxtaposition of last occurrences implies a hiatus at this level.
The polarity chron interpretation at Site 985 (Fig. 7; Table 3) is not aided by biostratigraphic constraints because of the absence of useful biostratigraphic datums at this site (Shipboard Scientific Party, 1996c). Natural gamma data from the shipboard multisensor track provide a means of correlating Site 985 with Site 907 (Fig. 10), providing corroboration for the magnetostratigraphic interpretations down to the base Gilbert Chron. The age-depth plot based on these interpretations indicates fairly constant sedimentation rates to ~160 mbsf (Fig. 11). No magnetostratigraphic interpretation is possible below this level because of drilling-related deformation of the core.