Detailed descriptive data are provided on the free-operant avoidance behavior of rats in a shuttlebox during both acquisition and terminal performance. Initially, eighteen 21-min acquisition sessions were given. Each hurdle-cross postponed the next shock 20 sec (response-shock interval) and shocks were scheduled every 5 sec (shock-shock interval) in the absence of a response. All eight subjects avoided over 70% of the shocks due (12/nmin) in Session 1. Maximum response rates were reached by the third session and declined slowly while shock rates continued to drop slowly through Session 15. Three subjects were run an additional five months with a response-shock interval of 20 sec and their terminal response rates were all under five responses per minute and shock rates were 0.07 per minute. Interresponse time distributions for terminal performance showed that over 95% of all responding by all three subjects occurred in the last half of the response-shock interval.Historically, the study of avoidance behavior evolved into two fairly distinct methodologies or "schools". The first (signalled escapeavoidance) involved an exteroceptive warning stimulus (CS), an escape contingency, some type of discrete trial procedure, and a shuttlebox or similar apparatus requiring a running response.In the second approach (unsignalled avoidance), no exteroceptive stimulus was employed and the trials feature was dropped for the free response situation. In the original report (Sidman, 1953) of this paradigm (unsignalled avoidance), brief, fixed-duration, inescapable shocks were scheduled at regular intervals in the absence of a response (S-S interval) and another shock-free period of time was produced by a response (R-S interval). The apparatus required a lever-press response.In both of these approaches, problems making experimentation difficult are commonplace. These may be unique to one approach, such as the lever-holding difficulties encountered in unsignalled bar-press avoidance (Hurwitz, 1967), or the long-term deterioration in performance ("decrements") often found in signalled shuttlebox escape-avoidance (Anderson and Nakamura, 1964). Other problems "Reprints may be obtained from Dave Riess, Research Hospital, Galesburg, Illinois 61401. frequently encountered are common to both approaches, such as inferior performance at the beginning of a session ("warm up effect"), as well as frequent failures to learn reported for both unsignalled bar-press avoidance (Weissman, 1962) and for signalled escapeavoidance (Brush, 1966).About 10 yr ago, Black and Morse (1961) conducted what was apparently the first study combining the free-operant paradigm with the more traditional shuttlebox apparatus, and their results with dogs were apparently successful in overcoming some of the common avoidance problems found in rats. Scobie (1970) has since combined the shuttlebox apparatus and the free-operant paradigm in successful avoidance conditioning with goldfish. The freeoperant shuttlebox combination has also been employed with rats with increas...