Two volcanic units were recovered at this site, of which only Unit 1 is suitable for radiometric dating (Duncan, Larsen, Allan, et al., 1996). Unit 1, a ~20-m-thick plagioclase-pyroxene-olivine-phyric basaltic lava flow with reverse magnetic polarity, has the highest K₂O content (0.41-0.47 wt%) and is the freshest volcanic unit sampled during Leg 163. The dating results of three whole rocks and one plagioclase separate yielded identical plateau ages within error (Table 1; Fig. 3). The weighted average of 49.6 ± 0.2 Ma is taken as the best estimate of age and uncertainty and is correlative with eruption during C22r (50.8-49.7 Ma).

The two volcanic units with normal magnetic polarity recovered at this site include an unusually thick (>69 m) compound flow (Unit 1) and the upper portion of the underlying flow (Unit 2) (Duncan, Larsen, Allan, et al., 1996). Two whole-rock samples from the massive, very fine-grained, aphyric lava of Unit 1 with low potassium content (0.24-0.29 wt% K₂O) yielded indistinguishable plateau ages within error (Table 1; Fig. 3); the weighted average age is 57.1 ± 1.3 Ma. Dating of the massive, sparsely to moderately plagioclase-phyric Unit 2 failed because of a lack in a consistent plateau. The normal magnetic polarity of these lavas (Hooper et al., Chap. 10, this volume) implies correlation with C26n (57.9-57.6 Ma) or C25n (56.4-55.9 Ma).

The 13 igneous flow units recovered at this site record a magnetic reversal from the reversely magnetized flow Units 3-13, at the base of the core, to normal magnetization, within the uppermost two flow units (Units 1 and 2) (Duncan, Larsen, Allan, et al., 1996; Hooper et al., Chap. 10, this volume). Two whole-rock samples from the moderately potassic (0.34-0.42 wt% K₂O), plagioclase-pyroxene-olivine-phyric basalt of Unit 1 yielded identical plateau ages within error (Table 1; Fig. 3) with a weighted average of 55.6 ± 0.6 Ma. This age is consistent with eruption during C25n (56.4-55.9 Ma). We note the next normally magnetized time intervals on either side of C25n, namely C24n.3n (53.3-52.9 Ma) and C26n (57.9-57.6 Ma), both deviate by >1 m.y. from the age range permitted by the 1 analytical uncertainty. Alternatively, however, it is possible that the normally magnetized lavas formed during a cryptochron that reflects a short excursion to normal polarization within a reversely magnetized magnetochron (Cande and Kent, 1992). Cryptochrons are plenty in C24r (e.g., C24r-11 at 55.574-55.565 Ma) (Cande and Kent, 1995). If these minor excursions in the magnetic field can really result in a complete change of the polarity with high inclination and high intensity, which is the case for the normally magnetized lavas in Units 1 and 2 of Hole 990A (Duncan, Larsen, Allan, et al., 1996; Hooper et al., Chap. 10, this volume), then the correlation of these lavas with one of the earliest cryptochrons in C24r (e.g., C24r-11) is an equal alternative to C25n. We note that the 9000-yr-long C24r-11 (Cande and Kent, 1995) easily permits the emplacement of two lava flows with an apparent average eruption frequency of ~670 yr (Larsen and Saunders, 1998). It is also noticeable that correlation with C24r-11 would fit the radiometric data strikingly well, although this precision really is beyond the resolution of argon dating in early Tertiary rocks.

For the underlying, reversely magnetized aphyric to highly olivine or plagioclase-augite-olivine phyric lava succession, four whole rocks and one plagioclase separate were analyzed from moderately potassic (0.33-0.41 wt% K₂O) samples. The isochron age for the plagioclase separate (plateau not developed) and three whole-rock plateau ages (representing Units 4, 5, and 7) yielded identical ages within error (Table 1; Fig. 3). The weighted average of 55.8 ± 0.6 Ma is taken as the best estimate of age and uncertainty for these lavas. Their reverse magnetization and stratigraphic position suggest correlation with C25r (57.6-56.4 Ma). Alternatively, if the normally polarized Units 1 and 2 are correlated to cryptochron C24r-11, the reversely lava Units 3-13 could have erupted during the earliest portion of C24r between 55.90-55.57 Ma, in good agreement with the radiometric data. The weighted average of all analyzed samples from Hole 990A is 55.7 ± 0.5 Ma, close to the boundary between C25 and C24 at 55.9 Ma.