The upper-shelf toughness regime corresponds to the normal operating temperature for ferritic nuclear pressure vessels. However, actual fibrous initiation toughness data have not been available for a wide variety of heats of steels and weldments. In fact, much of the data available have assumed that fracture initiation occurred at maximum load (equivalent energy approach), which is not the case for most materials tested on the upper shelf. Multispecimen (25.4 mm thickness compact or bend) initiation tests were, therefore, performed in the upper shelf temperature region to determine J_Ic for five base metal (two A533B-1, one A508-2, and two A302B) and four weld metal (two manual arc and two submerged arc) heats of nuclear ferritic steel. All of the heats were investigated under quasi-static loading (loading time ∼ 100 s) and under either closed-loop hydraulic loading (loading time ∼ 100 ms) or controlled impact loading (loading time ∼ 1 ms). For all materials investigated, crack initiation occurred prior to maximum load, and initiation toughness values (K_Jc) were up to 50 percent less than the equivalent energy (maximum load) toughness values. The initiation toughness results at 177°C also increased with decreasing loading time for all but one heat tested. HSST Plate 02 (A533B-1 steel) was tested at 71°C using all three loading rates; there were little differences observed for the Plate 02 tests due to loading rate, although the intermediate loading rate results were slightly lower than the rest. The data were analyzed using a simple statistical approach to obtain approximate confidence limits for the J_Ic values obtained. A comparison of J_Ic values obtained using three-point and nine-point fibrous crack length averages across the specimen thickness was made. The nine-point average generally gave higher values of J_Ic, but the variance about the regression line was not consistently lower than for the three-point average.