Two schemes of message storage were evaluated in a communications‐net simulation. An operator received messages, which frequently overlapped, from talkers on four channels. In a typical test problem, 32 relevant and six irrelevant messages were distributed over a five‐minute period. The operator's task was (1) to transcribe message information onto a plotting board and (2) to synthesize a problem solution from this information. In one scheme, Fixed Delay, the operator attempted to hear each message individually by appropriate use of a 9‐sec delay alternate circuit in each channel. In the other scheme, Variable Delay, messages were never heard directly; instead, message arrivals were indicated by lights and the operator summoned each message from storage individually. A third arrangement, in which all messages were heard directly as programed, served as a control condition. Performance criteria were speed, accuracy, and the number of repeats the operator had to request. The results for the Fixed Delay and control conditions were roughly equivalent. Under the Variable Delay condition more transcription and synthesis were complete and correct by the end of a programed sequence; only about half as much additional time and a third as many repeats were needed for problem solution.