Exerc., Vol. 42, No. 2, pp. 226–232, 2010. Marathon races faced with unexpectedly hot conditions must make a decision to start or not to start. The current race cancellation parameters may not reflect the safety profile of an individual race, and a universal temperature may not work an individual race. A retrospective review of the number of starters, finishers, and finish line medical encounters was used to calculate the rate of ‘‘unsuccessful’’marathon starters from race records, where unsuccessful is defined as medical encounters seen for any reason in the finish medical area plus marathon starters who did not finish. Unsuccessful marathon starters were plotted against the race start wet bulb globe temperature (WBGT), and the WBGT level at which a mass casualty incident (MCI) occurred was used to calculate a ‘‘do not start’’WBGT. A start WBGT of 921-C resulted in MCI or midrace cancellation in several races. Twin Cities Marathon data show a rapid increase in the rate of unsuccessful marathoner starters above a start WBGT of 13-C. The event experienced an area-wide MCI at a start WBGT of 22-C with an unsuccessful starter rate of 160 per 1000 finishers. Marathons in northern latitudes (940-) held in ‘‘unexpectedly’’hot conditions when the participants are not acclimatized and the start WBGT is 921-C often end in either race cancellation or an MCI. It would seem prudent not to start these races in similar conditions. The rate of unsuccessful marathon starters per 1000 marathon finishers plotted against start WBGT generates a curve that can be used to estimate a do not start level. The do not start WBGT for Twin Cities Marathon is 20.5-C on the basis of this model. Key Words: EXERTIONAL HEAT STROKE, HEAT STRESS, MASS CASUALTY INCIDENT, RACE