Recently reported radioisotopic dates and magnetic anomaly spacings have made it evident that modification is required for the age calibrations for the geomagnetic polarity timescale of Cande and Kent (1992) at the Cretaceous/Paleogene boundary and in the Pliocene. An adjusted geomagnetic reversal chronology for the Late Cretaceous and Cenozoic is presented that is consistent with astrochronology in the Pleistocene and Pliocene and with a new timescale for the Mesozoic. The age of 66 Ma for the Cretaceous/Paleogene (K/P) boundary used for calibration in the geomagnetic polarity timescale of Cande and Kent [1992] (hereinafter referred to as CK92) was supported by high precision laser fusion Ar/Ar sanidine single crystal dates from nonmarine strata in Montana. However, these age determinations are now considered to be anomalously old due to problems with sample preparation [Swisher et al., 1992, 1993]. A consensus is developing for an age of 65 Ma for the K/P boundary as recorded in marine [Swisher et al., 1992; Dalryrnple et al., 1993] and nonmarine [Swisher et al., 1993] sediments, and the 65 Ma age has been adopted, for example, as an anchor point in the Mesozoic timescale of Gradstein et al. [1994]. Astrochronologic control for the geomagnetic polarity timescale has been developed by Shackleton et al. [1990] and Hilgen [1991] for the Pleistocene and Pliocene to the base of the Thvera polarity subchron (subchron C3n.4n) and has been confirmed to about 3.3 Ma using high-precision At/At dating [Renne et al., 1993]. The astronochronologic estimates for the Brunhes/Matuyama (0.78 Ma) and Matuyama/Gauss (2.60 Ma) boundaries were already used for calibration in CK92; thus the good agreement of CK92 with the astronomical timescale to the older end of chron C2A (Gauss/Gilbert boundary) is not unexpected. An appreciable discrepancy, however, emerges in the early Pliocene where the astronomical timescale gives ages for the constituent polarity intervals of chron C3n (C3n. ln, C3n.2n, C3n.3n, and C3n.4n, or Cochiti, Nunivak, Sidufjall, and Thvera subchrons, respectively) that are systematically 150 to 180 kyr older than the interpolated ages in CK92. Wilson [1993] showed that the astrochronology gives a more consistent seafloor spreading history when applied to his revised spacings of anomalies on several Pacific spreading ridges. This points to the magnetic anomaly spacings for this interval used for interpolation by Cande and Kent [1992] as the likely source of the discrepancy and suggests that the available astronochronology provides reliable ages for polarity chrons through the Pliocene (see also Renne et al.,
1994).Copyfight 1995 by tho Amorican Ooophysical Union.