Recent publications have indicated that a KCl-stimulated ATPase from cereal roots is specifically associated with plasmalemma-enriched membrane fractions. However, in previous work we found that relatively high specific activities of this enzyme were also associated with a membrane fraction which did not contain plasmalemma. In an attempt to clarify this discrepancy, we have investigated the effect of density gradient composition upon the association of the enzyme with different membrane fractions isolated from the roots of Zea (10) also found a small peak of KCIstimulated ATPase activity from oat roots at low buoyant densities but regarded it as less significant than the activity associated with the plasmalemma.Because there were obvious discrepancies in the reported distributions of KCl-stimulated ATPase activities in membrane fractions from cereal roots, we felt that the problem merited further investigation. One possible explanation for the anomalies lay in the different centrifugation systems employed. Williamson and Wyn Jones (24) used a Ficoll gradient followed by a sucrose gradient, whereas Hodges and colleagues (6, 10, 12) employed only sucrose gradients. The observation that the composition of density gradients can markedly affect the buoyant density of membrane vesicles (2,15,20)