Developments in the genetic manipulation of mice have intensified interest in the relation between genes, environment and learned behavior, which in turn has led to exploration of experimental procedures for assessing genetic influences on learning using methods such as response acquisition. The requirement for multiple experimental control groups in such genetic comparisons studies amplifies the need for reliable, instrument-directed, learning assessment paradigms. The purpose of the present experiment was to implement such a procedure in several successive phases, including pre-food basal rates of interaction with sensors in a chamber, food-hopper training, and a simple disk-baiting procedure that produced differentially food-reinforced disk pressing (FR1) and provided quantitative measures of learning in outbred (CD-1) and inbred (C57BL/6J, BALB/cJ) mice. Response acquisition performances were measured in terms of the number of correct food-producing responses (disk press to hopper entry in less than 5s). The three mouse types showed differences in several performance measures prior to response acquisition training. Pre-food basal performances did not predict subsequent rate of acquisition of the target response. On average, CD-1 mice met a criterion of 50 food-producing responses slightly, but not significantly, faster than BALB/cJ mice. The C57BL/6J mice took significantly longer to meet the learning criterion and had slower response rates, due to longer after-reinforcement pausing. Procedural differences (massed versus partial/distributed training) and reinforcement parameters (duration of access) may differentially affect different mouse types independently of genetic differences in response acquisition.