Subaerially erupted tholeiites at Hole 642E were never exposed to the high-temperature seawater circulation and al teration conditions that are found at subaqueous ridges. Alteration of Site 642 rocks is therefore the product of the in teraction of rocks and fluids at low temperatures. The alteration mineralogy can thus be used to provide information on the geochemical effects of low temperature circulation of seawater. Rubidium-strontium systematics of leached and unleached tholeiites and underlying, continentally-derived dacites reflect interactions with seawater in fractures and vesicular flow tops. The secondary mineral assemblage in the tholei ites consists mainly of smectite, accompanied in a few flows by the assemblage celadonite + calcite (± native Cu). Textural relationships suggest that smectites formed early and that celadonite + calcite, which are at least in part cogenetic, formed later than and partially at the expense of smectite. Smectite precipitation occurred under variable, but generally low, water/rock conditions. The smectites contain much lower concentrations of alkali elements than has been reported in seafloor basalts, and sequentially leached frac tions of smectite contain Sr that has not achieved isotopic equilibrium. 87 Sr/ 86 Sr results of the leaching experiments suggest that Sr was mostly derived from seawater during early periods of smectite precipitation. The basalt-like 87 Sr/ 86 Sr of the most readily exchangeable fraction seems to suggest a late period of exposure to very low water /rock. Smectite for mation may have primarily occurred in the interval between the nearly 58-Ma age given by the lower series dacites and the 54.5 ± 0.2 Ma model age given by a celadonite from the top of the tholeiitic section. The 54.5 ± 0.2 Ma Rb-Sr model age may be recording the timing of foundering of the Vriring Plateau. Celadonites precipitated in flows below the top of the tholeiitic section define a Rb-Sr isochron with a slope corresponding to an age of 24.3 ± 0.4 Ma. This isochron may be reflecting mixing effects due to long-term chemical interaction between seawater and basalts, in which case the age provides only a minimum for the timing of late alteration. Alternatively, inferrential arguments can be made that the 24.3 ± 0.4 isochron age reflects the timing of the late Oligocene-early Miocene erosional event that af fected the Norwegian-Greenland Sea. Correlation of 87 Sr/ 86 Sr and 1/Sr in calcites results in a two-component mixing model for late alteration products. One end-member of the mixing trend is Eocene or younger seawater. Strontium from the nonradiogenic endmember can not, however, have been derived directly from the basalts. Rather, the data suggest that Sr in the calcites is a mixture of Sr derived from seawater and from pre-existing smectites. For Site 642, the reaction involved can be generalized as smec tite + seawater ++ celadonite + calcite. The geochemical effects of this reaction include net gains of K and C0 2 by the secondary mineral assemblage. The gross similarity o...
