The medulla oblongata of the brains of 71 scrapie-suspect cases were routinely fixed in 10 per cent formal saline and assessed for vacuolation on HE-stained sections. A pool of fresh brain material was also dissected from each animal and extracts prepared for the routine detection of scrapie-associated fibrils by negative stain transmission electron microscopy. The remaining formaldehyde fixed medulla samples, which were not used for the histological examination, were coded and subjected to a pretreatment with sodium borohydride and then processed using the routine fibril detection procedure. Of the 71 samples tested 46 were considered positive by all three test procedures. Sixteen samples were negative for all three tests. Four samples were positive by histopathological examination and positive for fibrils using fresh tissue, but fibrils could not be detected in the fixed tissue preparations. Conversely, there were five fixed samples in which fibrils could be detected which were negative for the other two tests. The fibrils observed in fixed preparations were indistinguishable from those observed in fresh tissue extracts. The sensitivity of the test for fibril detection using fixed tissue was 92 per cent and the specificity 76 per cent. It is concluded that scrapie-associated fibrils can be recovered from formaldehyde fixed tissue, as presented for routine histopathological examination, and therefore the method has potential in the retrospective analysis of archived brain tissue where only fixed material was stored.