One hundred and eighteen doctors who treat pulmonary neoplasms in Ontario were asked how they would wish to be treated if they had non-small cell lung cancer. Four different scenarios were given. The physicians were then asked if they would consent to take part as subjects in one or more clinical trials for which they would be eligible in those situations. The proportion of respondents who would consent to each study ranged from 11% to 64%. Reasons given for refusing to participate as subjects in each trial were varied, but many felt that the trials offered unacceptable options for treatment. Medical oncologists consented to each study more frequently than radiation oncologists, respirologists or thoracic surgeons but all disciplines ranked the 6 studies in the same order of acceptability. It is concluded that some patients with non-oat cell lung cancer currently receive experimental therapies with high risk/benefits ratios which experts in the field would not accept for themselves. It is suggested that the expert surrogate system may be useful as an adjunct to the institutional review board in evaluating new trials before they are activated.