A retrospective multiinstitutional study was initiated analyzing patients with T3 squamous cell carcinoma of the larynx, treated between 1975 and 1984 in eight Dutch head and neck cancer centers. In two institutions the treatment modalities used depended on the assessment of the response of the primary tumor to a first series of radiation. According to this split-course regime protocol, poor responders were offered surgery, that is, subsequent to a first series of radiation, a laryngectomy was performed (RT + S). For good responders (RT-I) or poor responders but medically unfit for and/or refusing surgery (RT-II), the radiation was continued after variable rest periods had elapsed (split-course radiation therapy). This paper analyzes 131 patients treated in the two institutions by this split-course regime protocol. Mean follow-up was 84 months; 89% were male, 11% female. At 10 years a corrected survival of 55% and an overall survival of 20% was observed. For the RT-I the actuarial local relapse rate was 56% versus a relapse rate of only 13% for the RT + S group. The corrected survival rates of the RT-I and RT + S were 54% and 70%, respectively. In contrast: 10 of 12 RT-II patients relapsed locally, and a corrected survival of only 21% was observed. The influence of overall treatment time (OTT) was studied by using the equation of the biological effective dose (BED), being defined as D(1 + d/alpha/beta) - (ln 2/alpha*Tp)*OTT). However, no clear relationship between BED and the risk for local and/or regional relapse could be established.