[en] Tensile results on austenitic stainless steels were analyzed to develop means for predicting strains at stresses above yield for reactor regulatory applications. Eight heats each of types 316 and 304 were tested at 24, 93, 204, and 316⁰C as mill-annealed and at 24⁰C after reannealing. The effects of heat-to-heat variations on total strain (to 5%) at discrete stress levels were portrayed by a rational polynomial incorporating three constants that relate to the basic features of the true-stress-true-strain diagram. Because these constants usually are interrelated, a single parameter, yield strength (YS), proved adequate to predict results. For predictions analytical expressions of yield strength, an average value (YSa), and a lower bound value [YSa - 1.65SEE (standard error of estimate)] were used. Using the rational polynomial with these parameters we determined (1) limits of total maximum strain and (2) ratios of strain of material of lower bound YS to that of average YS. These are recorded at regular increments of stress [34 MPa (5 ksi)] and at ASME Code-related stresses (S/sub y), S/sub m/, 1.2S/sub m/ and 1.5S/sub m/). At intermediate stresses, strain penalties for using material of lower bound strength were large, generally larger for type 316 than type 304. For mill-annealed type 316 at 24, 93, 204, and 316⁰C, the maximum ratios of strain were 8.8, 13.0, 14.1, and 14.9, respectively, whereas for type 304 they were 3.5, 3.4, 5.6, and 4.6. At 1.5S/sub m/ and 316⁰C, a maximum strain of 2.08% was predicted for type 316 and 1.66% for type 304, as contrasted to values of 0.14 and 0.39% for average strain