As long as the oxygen concentration at the sediment surface does not fall below 1 or 2 mg/liter, typical deep‐water sediments in the Great Lakes will probably be found to exert a measurable but quantitatively unimportant influence on the chemistry of the overlying waters. This conclusion is supported by the case of Windermere.The sequence of more conspicuous changes that take place when the oxygen does fall below 1 mg/liter at the interface is illustrated by the case of Esthwaite Water, representative of events in lakes where biological production or organic pollution is high, where the subthermocline volume is relatively small, or both. A progressive decline in oxygen concentration from 2 mg/liter to analytical zero at the interface was accompanied by a fall in electrode potential in the upper few millimeters of sediment, correlated with mobilization and transfer into the water first of manganese and later of iron. There is a concurrent transfer into the water of substantial quantities of phosphate, previously held in complex form, which may have important biological consequences. Other changes include liberation into the water of ammonia and silicate. Further reduction of the water‐sediment system permits microbial reduction of sulfate.A third example of sediment‐water exchange occurs during winter stratification under ice cover. In Torneträsk, in the course of 95 days of ice cover, a relatively thin layer of bottom water, initially at 1C, gained heat and bicarbonate from and lost oxygen to the sediments. The density increase, arising from the heat gain, set density currents in motion that drained this contact water into the deepest parts of topographically isolated subbasins.