Since its original description, differential display PCR (DD-PCR) has been extensively used in attempts to identify novel genes under a variety of circumstances. Despite its widespread use, however, few novel genes of interest have been identified. In the present study we describe a set of experiments examining reasons for failure of differential display. Evidence is presented that aberrant priming at both the 5' and 3' ends results in competition in the PCR, precluding detection of messages other than those which are abundantly expressed. Appropriate calculations are discussed which indicate this was predictable and unlikely to be overcome. While DD may be successfully applied in some settings, the evidence indicates that only abundantly expressed messages can be detected. This limitation is emphasized.