(Table 1) Helium concentration and flux in DSDP Hole 94-607 samples

Most of the helium-3 in oceanic sediments conies from interplanetary dust particles (IDPs), and can therefore be used to infer the accretion rate of dust to the Earth through time (Ozima et al., 1984, doi:10.1038/311448a0; Takayanagi and Ozima, 1987, doi:10.1029/JB092iB12p12531; Farley, 1995, doi:10.1038/376153a0). 3He records from slowly accumulating pelagic clays indicate that the accretion rate varies considerably over millions of years, probably owing to cometary and asteroidal break-up events3. Muller and MacDonald have proposed (Muller and MacDonald, 1995, doi:10.1038/377107b0) that periodic changes in this accretion rate due to a previously unrecognized 100-kyr periodicity in the Earth's orbital inclination might account for the prominence of this frequency in climate records of the past million years (Imbrie et al., 1993, doi:10.1029/93PA02751). Here we report variations in the 3He flux to the sea floor that support this idea. We find that the flux recorded in rapidly accumulating Quaternary sediments from the Mid-Atlantic Ridge oscillates with a period of about 100 kyr. We cannot yet say, however, whether the 100-kyr climate cycle is a consequence of, a cause of, or an effect independent of these periodic changes in the rate of delivery of interplanetary dust to the sea floor.

Age estimates of Discoaster datums and corresponding isotope stages (Table 2)

High-resolution records (2 7 kyr) of Upper Pliocene Discoaster abundances obtained from six ODP/DSDP sites are assessed independently using oxygen isotope stratigraphy. Four Atlantic Ocean sites (DSDP Sites 552 and 607, and ODP Sites 659 and 662) comprise a transect from 56°N to 1°S and provide a record of latitudinal variations in Diseoaster biogeography. Low-latitude sites in the Atlantic (ODP Site 662), Pacific (ODP Site 677), and Indian (ODP Site 709) oceans provide additional information about variability in Discoaster abundance patterns within the equatorial region. A common chronology, based on the astronomical time scale developed for ODP Site 677, has been established for all the sites. By integrating oxygen isotope data and Discoaster abundance records at each site we are able to independently evaluate the temporal and spatial distribution of D. brouweri and D. triradiatus in the 500 kyr prior to the extinction of the discoasters near the base of the Olduvai subchron. Major decreases in abundance are evident during some of the more intense late Pliocene glacial events. In particular, glacial isotope stages 82, 96, 98 and 100 are associated with distinct abundance minima. At these times, large-scale changes in surface hydrographic conditions appear to have suppressed Discoaster numbers on a global scale. The increase in abundance of D. triradiatus, which precedes the extinction of the discoasters by around 200 kyr, may also be related to the intensification of environmental pressures that accompanied the build-up of Northern Hemisphere ice sheets during the late Pliocene. In spite of contrasting geographic and oceanographic settings, the various D. brouweri and D. triradiatus records are remarkably similar. This demonstrates that the acme and extinction events are excellent biostratigraphic datums. The simultaneous extinction of D. brouweri and D. triradiatus at 1.95 Ma were synchronous events at both a regional scale within the Atlantic, and on a global scale between the three major oceans. However, the start of the D. triradiatus acme appears to have been diachronous, occurring some 40 kyr earlier in the Atlantic than in the Indo-Pacific, and hence the stratigraphic usefulness of this datum is regional rather than global.