To examine whether changes in the effective doubling time of tumor cells during irradiation of head-and-neck cancer are linked to accumulating dose.Optimal fitting of the results of four apparently iso-effective regimens in three recently reported randomized controlled trials (continuous hyperfractionated accelerated radiation therapy [CHART], Radiation Therapy Oncology Group [RTOG] 90-03, and Trans-Tasman Radiation Oncology Group [TROG] 91.01) was attempted using two different types of model of the change in effective doubling time that may occur during treatment. The first involved the traditional approach where doubling time changes at specific times after the start of treatment regardless of fractionation used (''fixed response time'' models). The second is where changes in doubling time are linked to accumulating biologic dose (''cellular depletion'' models).Changes in effective doubling time occur during radiotherapy for head-and-neck cancer. Data from the three trials can be fitted successfully by functions that imply a continuous reduction in effective doubling time. Models linking the reductions in effective doubling time to the cellular depletion that occurs in the tumor during radiotherapy fit the data satisfactorily. Effective doubling time ultimately reduces to 2 days, or slightly less, during high-dose radiotherapy regimens designed to cure squamous head-and-neck cancer.If the assumption of iso-effectivity is justified then this study indicates that the ''repopulation phenomenon'' may be best described by a function that depicts a reduction in effective tumor cell doubling time that decreases continuously during treatment down to a certain minimum value. Furthermore, this reduction may be linked to cellular depletion.