Three putative differentiation inducing agents and five conventional drugs which have been shown to be active in patients with squamous cell carcinoma of the head and neck (HNSCC), were tested for activity against HNSCC transplanted in nude mice. Drugs were administered at a maximum tolerated dose level. By testing the effect of conventional drugs the value of the nude mouse xenograft model for testing new drugs was assessed. Bleomycin as well as cisplatin showed antitumour effects in nude mice, although toxicity and thereby the effectiveness of cisplatin varied during the 5-year period in which the experiments were performed. Bleomycin caused responses in 4 out of 13 tumour lines and cisplatin, when administered at a high dose, was active in 2 out of 5 tumours. 5-Fluoroura-cil and cyclophosphamide were only moderately active, while methotrexate was inactive. These data indicate that the model might be of value in the detection of new anticancer drugs although it may have a tendency to underestimate the activity of some drugs. We used the nude mouse xenograft model to test the antitumour activity of three differentiation inducing agents, the polar-planar solvents hexamethylene bisacetamide (HMBA) and N, N-dimethylformamide (DMF) and the antimetabolite 5-aza-2'deoxycytidine (5-aza-dCyd). HMBA appeared to be inactive. In contrast, DMF was active in 1 out of 4 tumour lines while 5-aza-dCyd was active in 2 out of 5 lines tested. Furthermore, the whole panel of differentiation inducers and conventional drugs was tested for antitumour activity against three HNSCC tumour lines. The data indicate that in some human HNSSC tumours the differentiation inducing agents 5-aza-dCyd and DMF have a better antitumour activity than the conventional drugs. In contrast no antitumour activity of HMBA was found.