Differential display (DD) is a powerful molecular tool that allows the identification and subsequent isolation of transcripts differentially expressed between biological samples, for example, between undifferentiated and differentiated cells, between different tissues or in one tissue at different stages of development. However, significantly high rates of apparent false positives have been reported using this technique. We suggest that the vast majority of false positives do not represent the originally selected transcript, but instead result from the re-amplification of cDNA species that co-migrate with the cDNA of interest in DD gels. Here we describe the use of a procedure to resolve co-migrating cDNAs and to purify the candidate of interest before cloning. The use of this modified technique resolves downstream problems encountered during DD experiments.