Determination of reliable crystallization ages by K-Ar methods for young (<1 Ma), fresh basalts from the seafloor has been frustrated by several effects. The small amounts of radiogenic 40Ar developed over these timescales in such low-K rocks are difficult to resolve from predominantly atmospheric 40Ar. An additional concern is that mantle-derived 40Ar may not be totally outgassed when magmas quench at seafloor hydrostatic pressures. We have developed a successful strategy for partially separating atmospheric from radiogenic 40Ar in 40Ar-39Ar incremental heating experiments on mid-ocean ridge basalts, from the East Pacific Rise near the Clipperton Fracture Zone. Pre-heating samples to 400¿C removes surficial atmospheric 40Ar, while subsequent heating at 600¿--1000¿C produces favorable proportions of radiogenic 40Ar. Experimentals using ''zero-age'' glassy and holocrystalline samples show that mantle-derived 40Ar is retained only in the outermost few cm of rapidly cooled pillow basalts. ¿ American Geophysical Union 1994