Compared to rate of discrete responding, response duration has received scant attention as a dependent measure in behavior analysis. We will present results from a series of recent experiments in which we studied the duration of lever-pressing in laboratory rats. In all of the investigations, we arranged the delivery of edible reinforcers if depressing a response lever fell under or exceeded a programmed interval. We will discuss three general findings in detail. First, response duration may present some advantages over response rate for studying the effects of delayed reinforcement. Second, response duration increases significantly during extinction. This occurs whether or not there has been a history of differential reinforcement for depressing the response lever. This has implications for the interpretation of resurgence when behavior has a salient duration dimension. And finally, arranging reinforcement for response duration results in chains of idiosyncratic responses that rats emit concurrently with lever pressing. Whether these responses should be operationally defined as lever pressing presents a conceptual challenge.