Tumor clonogenic cell content is believed to play an important role in the outcome of radiotherapy. However, there is no proven method to assess the number of clonogens in human tumors accurately. All currently available assays employ in vitro plating efficiency or in vivo TD50 (the average number of cells needed to induce tumors in 50% of injected mice) to estimate the tumor clonogenic ability. In this study, a monolayer mass primary culture system was used to estimate the clonogenic cell fraction in human tumors. For this purpose, 25 growth curves were performed for 25 tumor specimens derived from 21 head and neck and 4 cervical squamous cell carcinomas. The exponential portion of each growth curve was extrapolated through the ordinate (day 0) to estimate the clonogenic cell fraction; this method is only an estimate because it assumes no lag phase before exponential growth of clonogenic cells. The mean clonogenic cell fraction, expressed as clonogens/tumor cells inoculated, was relatively low (mean: 0.71%, range: 0.11-9.28), and the variation was wide (coefficient of variation = 148%). On the other hand, the doubling time of the growing population was 1.46 days and exhibited a very narrow range (0.98-2.24, coefficient of variation = 24%). The mean and range of clonogenic cell fraction were found to be in agreement with published values of soft agar colony forming efficiencies in both murine and human tumors. However, further investigation is necessary to determine how accurately this method measures the relative clonogenic cell content in human tumors. Clinical correlations between clonogenic cell fraction values and the response to radiotherapy are still too early to determine.