It has previously been observed that Cl−influx falls with increasing internal CI−concentration. This paper attempts to show that this is a bona fide negative feedback involved in a homeostatic CI−accumulation system. First, predictions are made of how the final steady concentration to which Cl−is accumulated ([Cl]1∞) would be expected to vary as external Cl−concentration ([Cl]0) or temperature, change. This is done for the cases where influx has (a) no feedback control, or is (b) under error-actuated or (c) under reciprocal feedback control. Secondly, Cl−influx and [Cl]1∞are measured in carrot and maize root tissue over a range of [Cl]o and temperatures for comparison with the prediction. [Cl)1∞varies by only about 20% over a range of [Cl]0 and temperatures which have large immediate effects on Cl−influx. This is consistent only with error-actuated feedback control. Changes in [Cl]0 and temperature are best regarded simply as perturbations of Cl−influx since they have no marked effect on [Cl)1∞. In such a homeostatic system the feedback signal is a function of (CR – [Cl], ), where CR represents the value of an inbuilt set point. CR must be constant with [Cl]o and temperature, and its nature is briefly discussed. It is further predicted that the accumulated level can only be altered in a controlled manner by a change in CR, and that this would alter influx by the same proportion. The effects of abscisic acid in increasing both [Cl]1∞ and Cl−influx are consistent with an effect primarily on CR. These conclusions imply that when varieties or species differ in levels accumulated, they do so not because influx isotherms differ but rather because their set points differ.