Stream sediments adsorb certain solutes from streams, thereby significantly changing the solute composition; but little is known about the details and rates of these adsorptive processes. To investigate such processes, a 24-hr. injection of a solution containing chloride, strontium, potassium, sodium and lead was made at the head of a 640-m reach of Uvas Creek in west-central Santa Clara County, California. Uvas Creek is a cobble-bed pool-and-riffle stream draining the eastern slopes of the Santa Cruz Mountains. By September 12, 1973, after a long dry season, Uvas Creek had a low (0.0215 m3s-1 average) flow which varied diurnally, from 0.018 to 0.025 m3s-1. Because stream discharge varied while the injection rate was constant, the concentration of tracers (injected solutes), after mixing in the stream, varied inversely with discharge. Chloride, a nonreactive solute, served as a tracer of water movement. Analysis of extensive chloride concentration data at five sites below the injection point during and after the injection demonstrated that there was considerable underflow of water through the stream gravels; however, the extent of underflow varied greatly within the study reach. Pre-injection water, displaced by tracer-laden water percolating through the gravels, diluted tracers in the stream channel, giving the mistaken impression of groundwater inflow at some points. Accurate measurement of total discharge in such streams requires prolonged tracer injection unless a reach can be found where underflow is negligible. Strontium and potassium were adsorbed by the bed sediments to a moderate extent and lead was strongly adsorbed. A high proportion of these metals could be removed by adsorption from percolating underflow because of extensive and intimate contact with bed sediments. After channel clearing following injection cutoff, 51% of the added strontium and 96% of the lead remained in the study reach, whereas only 19% of the chloride remained. Packets of sized sediment, placed in the stream before the experiment and withdrawn during and after the injection, indicated that the strontium absorbed on the 0.42-0.50-mm size sediment appeared to achieve near equilibrium with dissolved strontium within less than 2 hr. whereas 3.4-4.0-mm grains had not reached that stage after 24 hr. The cation-exchange capacity (CEC) of the sediments shows a "bimodal" distribution with grain size. Largest values are in the finest sizes, lower values in the fine-to-medium sand-size range, intermediate values in the coarse- to very coarse-grained sand, and decreasing values with size above very coarse-grained sand. This considerable exchange capacity in coarse-sand to granule-size particles means that a streambed, that has not been infilled with fines to reduce permeability, can be highly reactive and accessible throughout a rather thick sediment layer and hence have a large and available reactive capacity. As stream discharge increases from low flow, the ratio of underflow to channel flow should decrease rapidly with resultant diminution in percent of solutes sorbed within a particular stream reach. ?? 1984.