Product tests are generally intended to serve two purposes, first to expose those mechanisms which are likely to prove troublesome so that they may be eliminated and second to provide estimat)es of the performance parameters of the product. In this paper, systems in which the failure rate is the significant performance parameter and in which failures constitute a Poisson process are considered. It is assumed that every system failure may be classified as “correctable” or “intrinsic” and that every correctable failure which occurs during test leads to elimination of the mechanisms which caused it. Optimal test plans are developed based on the purposes of the test and the economic factors of importance to the manufacturer. © 1965 Taylor & Francis Group, LLC.