This report documents the error rate in a subset of the IMAGE Consortium mouse cDNA library. A set of 1189 stocks were purchased from and resequenced prior to the construction of a cDNA microarray to be used in toxicogenomic studies of endocrine disruption in the mouse testis. Of the clones purchased, only 62.2% (740) matched their published sequence. Using an agarose gel electrophoresis prescreening strategy, 362 stock cultures showed the possible presence of two or more unique plasmid species. After isolation of individual colonies from these stocks, it was determined that 7.1% of the original 1189 stocks contained both a correct and an incorrect plasmid. An additional 6.7% of the original 1189 stocks contained multiple, distinct, incorrect plasmids, indicating the likelihood of multiple contaminating events. 330 contaminating or incorrect sequences could be assigned to unique gene clusters in the murine UniGene database based on similarity to murine sequences in dbEST. An additional 77 of the contaminating or incorrect clones had significant identity to a murine sequence in dbEST. While only 740 of the stocks purchased contained the desired cDNA clone, agarose gel prescreening, colony isolation, and similarity searching of dbEST allowed for the identification of an additional 407 clones that would have otherwise been discarded. Some stocks yielded sequence which matched published sequence information, yet did so with low identity and high ambiguity. Several of these stocks were subsequently discovered to contain multiple plasmid species, indicating the need for stringent sequence verification criteria. Rapid, high throughput assays of suspected chemical toxicants using cDNA microarrays will be greatly assisted by the use of high quality cDNA clones, thus minimizing false positive results and maximizing the number of informative data points. Considering the high error rate in this subset of the IMAGE cDNA library, the use of sequence verified clones or prescreening non sequence verified clones with agarose gel electrophoresis is warranted prior to constructing microarrays.