Injection of ethyl alcohol in high concentrations into tissues produces coagula-tive necrosis. The benefits of direct injection of 98 % ethanol into hepatic metastases from 1,2-dimethylhydrazine (DMH)-induced colonic cancer was investigated in groups of 20 Sprague-Dawley rats of either sex. At 10 weeks of age, rats were subcutaneously injected every week with 10 mg/kg DMH for 28 weeks. They were then housed for 3 months. At the end of this period all the animals had developed colonic carcinoma with multiple hepatic metastases. Total colectomy and the fashioning of an ileostomy coupled with direct injection of 0.1–0.2 ml of ethanol into each of the hepatic metastases, after mobilizing the liver by dividing its fascial attachments to facilitate easier tumour detection by inspection and palpation, afforded a significant survival advantage relative to colectomy alone (20.1 + 0.2 months of age, vs, 12.8 ± 0.2 months of age, mean + SEM, n = 20, p < 0.01 by the Mann-Whitney U test). The clinical implications of these observations were, therefore, examined in a pilot study carried out on 6 patients (2 women and 4 men with an age range of 43–71 years, mean 56) who had adenocarcinoma of the sigmoid colon with multiple hepatic secondaries. The colonic tumour was resected and an end-to-end anastomosis effected, then all palpable hepatic metastases were slowly injected with 1–1.5 ml of 98% ethanol. Two weeks post-operatively and thereafter once every 2 months any hepatic lesions detected ultrasonically were similarly treated per-cutaneously. This treatment produced no adverse effects of any significance. None of the patients died during the first post-operative year. Death was due to disease spread in all the patients. The mean survival measured from the time of tumour resection until death from any cause was 20 months (range 17-26 months). It thus appears that chemonecrosis for the treatment of liver metastases from colonic cancer is a simple and safe therapeutic modality which offers a surviva