When a mantle plume interacts with a mid-ocean ridge, both are noticeably affected. The mid-ocean ridge can display anomalously shallow bathymetry, excess volcanism, thickened crust, asymmetric sea-floor spreading and a plume component in the composition of the ridge basalts (Schilling, 1973, doi:10.1038/242565a0; Verma et al., 1983, doi:10.1038/306654a0; Ito and Lin, 1995, doi:10.1130/0091-7613(1995)0232.3.CO;2; Müller et al., 1998, doi:10.1038/24850). The hotspot-related volcanism can be drawn closer to the ridge, and its geochemical composition can also be affected (Ito and Lin, 1995, doi:10.1130/0091-7613(1995)0232.3.CO;2; White et al., 1993, doi:10.1029/93JB02018; Kincaid et al., 1995, doi:10.1038/376758a0; Kingsley and Schilling, 1998, doi:10.1029/98JB01496 ). Here we present Sr–Nd–Pb isotopic analyses of samples from the next-to-oldest seamount in the Hawaiian hotspot track, the Detroit seamount at 51° N, which show that, 81 Myr ago, the Hawaiian hotspot produced volcanism with an isotopic signature indistinguishable from mid-ocean ridge basalt. This composition is unprecedented in the known volcanism from the Hawaiian hotspot, but is consistent with the interpretation from plate reconstructions (Mammerickx and Sharman, 1988, doi:10.1029/JB093iB04p03009) that the hotspot was located close to a mid-ocean ridge about 80 Myr ago. As the rising mantle plume encountered the hot, low-viscosity asthenosphere and hot, thin lithosphere near the spreading centre, it appears to have entrained enough of the isotopically depleted upper mantle to overwhelm the chemical characteristics of the plume itself. The Hawaiian hotspot thus joins the growing list of hotspots that have interacted with a rift early in their history.

All isotopic ratios were measured at Cornell University. All samples were repeatedly leached in hot HCl before digestion, and all isotopic ratios were corrected for fractionation. Sr and Pb isotope ratios are normalized to standards NBS987 ( = 0.710248) and NBS981 ( = 16.937; 15.493; 36.705), respectively. The average 143Nd/144Nd for the La Jolla Nd standard measured at the time of these analyses (via Ames Nd) was 0.511876614. Two-sigma standard errors based on numerous analyses of standards during the time of the unknown runs are: 87Sr/86Sr ± 0.000014, 143Nd/144Nd ± 0.000016, 206Pb/204Pb ± 0.010, 207Pb/204Pb ± 0.011, and 208Pb/204Pb ± 0.030. Within-run standard errors are smaller than these external errors. Trace-element concentrations were measured by isotope dilution thermal ionization mass spectrometry (TIMS-ID) of leached samples (Pb, Th and U) and inductively coupled plasma mass spectrometry (ICP-MS), of unleached samples (Rb, Sr, La, Nd and Sm). The ages of the Site 433 and Site 884 samples are known from 40Ar±39Ar dating (Clagues and Dalrymple, 1987; Lanphere et al., 1980, doi:10.2973/dsdp.proc.55.131.1980). For age correction purposes, the Site 883 samples are assumed to be 1 Myr younger than the Site 884 samples because Site 883 was slightly higher on the seamount, and the Site 883 samples appear to be of transitional composition. The Site 192 sample is estimated to be 85 Myr old for age correction purposes. Errors in the estimated ages of the Site 192 and Site 883 samples of several million years or less are not relevant here.

Supplement to: Keller, Randall A; Fisk, Martin R; White, William M (2000): Isotopic evidence for Late Cretaceous plume-ridge interaction at the Hawaiian hotspot. Nature, 405(6787), 673-676