Modern river studies support the generalization that surface waters in regi
ons undergoing extensive chemical weathering have elevated dissolved german
ium/silicon (Ge/Si) ratios compared to regions with less extensive chemical
weathering, thought to be the result of Ce fractionation in 2:1 clays. Tem
poral variations in Ge/Si observed in marine diatoms have thus been linked
to past global trends in terrestrial weathering. However, this relationship
has not been adequately ground-truthed by a terrestrial-based paleoclimate
study utilizing Ge/Si ratios in lake diatoms. To this end, a sediment core
was extracted from Dry Lake (el. 2763 m) in the headwaters of the Santa An
a River of southern California in July 1996. The Dry Lake drainage basin is
comprised of biotite-muscovite gneiss and granite, with sparse pine forest
s and relatively high relief. The core had a basal AMS C-14 age of 8,350 +/
- 60 ybp. We successfully separated diatom samples large enough for cleanin
g, dissolution and chemical analysis without contamination by detrital mate
rials (confirmed by trace metal analyses). Two dissolved opal samples were
analyzed for Ge and Si concentrations along with modern water samples colle
cted from waters within the drainage basin. Diatoms obtained from 8000-year
-old sediments near the bottom of the core had an opal Ge/Si of 0.79 x 10(-
6) (mol/mol). A composite sample of diatoms deposited within the past 100 y
ears yielded a significantly lower Ge/Si of 0.34 x 10(-6). Analysis of Ge/S
i was also performed from a variety of other materials in the Dry Lake basi
n, including stream and lake water, unweathered bedrock, soils, and mineral
separates. Together with sedimentologic records from the lake sediments an
d Ge/Si recorded in other basin materials, it appears that the high Ge/Si v
alues recorded in the 8000-year opal sample were the result of preferential
early weathering of high Ge/Si biotite and muscovite minerals from the slo
pes. Thus, the detailed examination of Ge/Si cycling in this isolated basin
indicates that factors other than clay mineral transformations may drive G
e/Si paleorecords in some settings. (C) 2000 Elsevier Science B.V. All righ
ts reserved.