D(2)O is the only "chemical" agent that consistently affects the frequency of circadian oscillations: its effect is now known to be so widespread and predictable that its action merits closer study as a potential clue to the currently obscure concrete nature of circadian oscillators. The great diversity of D(2)O effects on biological systems in general is briefly reviewed and the need for rejectable hypotheses concerning the action of D(2)O on circadian clocks is stressed because current speculation on its action yields "predictions" expected from almost any hypothesis. We consider the hypothesis that it "diminishes the apparent temperature" of the cell and proceed to test this by examining the effect of D(2)O on temperature-dependent and temperature-compensated aspects of the circadian system in Drosophila. We find these components respond as differentially to D(2)O as they do to temperature; we conclude, however, with a warning that this result may be equivocal if, as we now suspect, the frequency of circadian oscillations is generally homeostatically conserved-not only in the face of temperature change, but change in any variable to which it is sensitive. More crucial tests of the temperature-equivalence hypothesis for D(2)O action are defined.