The terminal transferase uridyl nick end labelling (TUNEL) assay allows the easy demonstration of cell death as a result of apoptosis. However, when this assay is applied to liver tissue, the number of TUNEL positive cells is dependent on the time of incubation with proteinase K. To test whether false positive results are the result of the release of endogenous endonucleases by proteinase K and can be abolished by pretreatment with diethyl pyrocarbonate (DEPC). Involution of hyperplastic ductules in bile duct ligated rats after biliary decompression by Roux-en-Y anastomosis and acute CCl4 intoxication were studied as models of apoptosis and necrosis, respectively. A standard TUNEL assay was applied to formalin fixed tissue sections mounted with cement. To inhibit putative endogenous endonucleases, tissue slides were pre-incubated with DEPC. In the standard TUNEL assay, the number of positive nuclei was highly dependent upon the length of time that sections were incubated with proteinase K. After pretreatment with DEPC, only cells that also exhibited morphological features of apoptosis stained positive. DEPC pretreatment abolished false positive staining in CCl4 induced hepatocyte necrosis and blocked interference by endogenous alkaline phosphatase in intestine. The method of gluing the tissue section to the glass slide was found to be of utmost importance because the effect of DEPC was abolished on silanised slides. False positive staining in the TUNEL assay in the liver is caused by the release of endogenous endonucleases as a result of proteinase treatment. This can be abolished by pretreatment of tissue slides with DEPC.