In certain areas, supplies have been furnished to storage water-heaters and other such loads through relays permitting the supply to be temporarily suspended by superimposed control during peak times at the discretion of the engineer. The tests described in this paper were carried out in two areas in which such methods of control were available. The tests were made independently of peak-time suspension, and were aimed at finding out how many heaters were consuming at any given time. For this purpose the supply was arbitrarily cut off from the heaters and resumed again as quickly as possible. The changes in the feeder load were then used as a measure of the heater load at that moment. By repeating the process at regular intervals it was possible to separate out the heater load curve from the curve for the remainder of the load, and thus determine the diversity of the water-heater load.