A TEST OF (GE SI)(OPAL) AS A PALEORECORDER OF (GE/SI)(SEAWATER)/

Late Pleistocene variations of germanium to silicon ratios in marine d iatom shells from sediment cores, (Ge/Si)(opal), are coherent with the global isotope record of glacial to interglacial climate change. Thes e variations are thought to reflect changes in (Ge/Si)(seawater) drive n by climate-modulated alterations in oceanic Ge/Si sources and sinks. However, an important criterion for interpreting (Ge/Si)(opal) as a m onitor of whole ocean (Ge/Si)(seawater) is that the opal burial ratio be insensitive both to local diatom production and surface ocean silic a concentrations (so-called biological fractionation effects) and to d ifferential dissolution artifacts (so-called diagenesis offsets). Here we test these assumptions by comparing model ocean sediment (Ge/Si)(o pal) distributions with data from Holocene and glacial sediments acros s the high-latitude Indian-Antarctic Ocean siliceous ooze belt. In con trast to the model, the data show no gradients in either Holocene or g lacial (Ge/Si)(opal) values across productivity zones displaying drama tic changes in biosiliceous production, opal burial, and dissolution. This evidence supports the contention that fractionation effects are s mall and that observed down-core variations in (Ge/Si)(opal) faithfull y record secular changes in (Ge/Si)(seawater).