The data show a wide range in relaxation strengths. Study of the data shows that the following factors are involved: 1. The major cause for the wide range in relaxation strengths at 850 F appears to be molybdenum content, as is shown by the figures on page 20. For the quenched-and-tempered condition, the relaxation strengths increased directly with molybdenum content up to 0.5 per cent over a comparatively narrow band. A rather wide range of carbon contents and heat-treating conditions as well as the variations in initial stress were definitely less influential than molybdenum content up to 0.5 per cent. There is some limiting molybdenum content for this relationship because 2.16 per cent gave no higher strength than 0.5 per cent. 2. There was some indication that relaxation strengths at 850 F tended to be slightly higher when the quenching temperature and initial stresses were increased or when the tempering temperature was lowered. There was no relation to tensile strength at room temperature, to carbon content, or to oil versus water quenching for the quenched-and-tempered materials. 3. Relatively soft 0.16 per cent carbon material produced by normalizing and tempering at 1200 F had low relaxation strengths in relation to the molybdenum content. 4. The data at other temperatures than 850 F were too sparse to warrant conclusions, except that when high tensile and yield strengths were obtained, by normalizing and restricting the tempering to temperatures below 1200 F, the relaxation strengths were also high. 5. The curves, drawn for variation in relaxation strength with temperature, were based on average values for molybdenum contents of 0.25 and 0.50 per cent at 850 F using the correlation on page 19. At other temperatures, the actual data were used, adjusted for variation in molybdenum content. The abnormally-low values for the normalized 0.16 carbon Material 2 were omitted in establishing the curves.