A method is proposed for determining a lower-bound fracture toughness value from the result of a single test in the transition region for steels. The method is based on the size criterion of Iwadate et al. for determining the number of tests needed to characterize toughness adequately at a given condition and a Weibull statistical distribution. The method was evaluated by determining a lower-bound toughness value for 68 sets of data, which include over 700 fracture toughness values in the transition. The lower-bound estimate was judged to be conservative if all of the lower-bound estimates were lower than the lowest toughness value in the set. The method was conservative for more than 60% of the data sets and for more than 95% of the individual toughness values. In addition, the method was evaluated by comparing it with 4 methods that used multiple specimens for determining a lower-bound toughness. The single-specimen method gave essentially the same lower-bound estimates as the other multiple-specimen methods.