Reaction times were collected for three observers in aYES‐NOsignal detection experiment, for two different signal strengths, and threea prioriprobabilities of a signal occurrence—P(S): 0.2, 0.5, and 0.8. Each server gave 1800 responds under each of the six conditions. A trial ended only when the observer terminated it. Some of the major findings are (1) latencies have Pearson distributions, mainly Beta of the first and second kind; (2) there is an apparent genetic character to the distributions—that is, they resemble each other according to the response (YESorNO) rather than the stimulus (signal or noise alone); (3) atP(S) = 0.2, the mean latency of a false rejection of a signal is least, that of a correct detection greatest, but forP(S) = 0.5, and 0.8, mean latency is least for correct detection, and greatest for correctNO'S, and (4) generally, reaction times are inversely related to thea prioriprobability of their associated stimulus classes.