|Abstract|| ||The conditions of networked ensemble performance were simulated in an experiment. Pairs of musicians were placed apart in isolated rooms and given a simple rhythm to clap together. A microphone was placed as close as possible to each performer's hands. Each monitored the other's sound via headphones, and a delay was introduced between the source and listener. Starting tempo, given by a recorded count-in, and delay time were varied across trials. Recordings of the trials were analyzed with a precise event detection algorithm to locate clap onsets, from which the tempo was inferred. The rate of deceleration increased with longer delays, while shorter delays (<= 11.5 ms) produced a modest, but significant acceleration. The goal is to identify the region of delay time that is most conducive to maintaining a steady tempo. This will help to determine the necessary delay conditions to support networked musical performance (which may be over long distances or in adjoining rooms). Humans performed significantly better than a simple model of a memoryless instantaneous reaction.