MILANKOVITCH THEORY VIEWED FROM DEVILS HOLE

VARIATIONS in the oxygen isotope content (deltaO-18) of late Quaternar y deep-sea sediments mainly reflect changes in continental ice mass1, and hence provide important information about the timing of past ice a ges. Because these sediments cannot yet be dated directly beyond the r ange of radiocarbon dating (40-50 kyr), ages for the deltaO-18 record have been generated2,3 by matching the phase of the changes in deltaO- 18 to that of variations in the Earth's precession and obliquity. Adop ting this timescale yields a close correspondence between the time-var ying amplitudes of these orbital variations and those of a wide range of climate proxies4, lending support to the Milankovitch theory that t he Earth's glacial-interglacial cycles are driven by orbital variation s. Recently Winograd et al.5 reported a record of deltaO-18 variations in a fresh-water carbonate sequence from Devils Hole, Nevada, dated b y U-Th disequilibrium6. They concluded that the timing of several of t he features in the record, which reflects changes in the temperature o f precipitation over Nevada as well as changes in the isotopic composi tion of the moisture source5,7, showed significant deviations from tha t predicted by Milankovitch theory. Here we demonstrate that applying the Devils Hole chronology to ocean cores requires physically implausi ble changes in sedimentation rate. Moreover, spectral analysis of the Devils Hole record shows clear evidence of orbital influence. We there fore conclude that transfer of the Devils Hole chronology to the marin e record is inappropriate, and that the evidence in favour of Milankov itch theory remains strong.