The investigational antitumour agent 5,6-dimethyl-xanthenone-4-acetic acid (5,6-MeXAA) induced dose-dependent haemorrhagic necrosis of colon 38 tumours to a similar extent to that induced using bacterial lipopolysaccharide (LPS). TNF-alpha activity in serum and mRNA for TNF-alpha in splenocytes were induced over a broad range of LPS doses, whereas with 5,6-MeXAA, induction occurred only at concentrations approaching the maximum tolerated dose. At concentrations that provided similar degrees of haemorrhagic necrosis, the levels of serum TNF-alpha induced using 5,6-MeXAA were 100-fold lower than those obtained with LPS, indicating that haemorrhagic necrosis was not directly correlated with TNF-alpha levels. There was also no correlation between the degree of tumour necrosis and the duration of growth delay. Treatment with LPS did not induce a significant delay in growth, despite extensive tumour haemorrhagic necrosis, whereas with 5,6-MeXAA, treatments that improved the cure rate did not necessarily give longer growth delays. Therapy using a combination of sub-optimal doses of both compounds was synergistic for the induction of serum TNF-alpha and message for TNF-alpha but was not synergistic for antitumour efficacy. Thus, no correlation is evident between cure rates, duration of growth delay, haemorrhagic necrosis and TNF-alpha induction by 5,6-MeXAA or LPS.