Eighteen commercial heats of Ni-Cr-Mo-V steel have been subjected to both step-cooling embrittling heat treatments and isothermal exposures at 750, 850, 950, and 1050 F for times up to 40,000 h. Embrittlement was measured in the as-heat treated condition on both rotor forgings and bucket wheel forgings, as well as in a deembrittled condition after water quenching from 1100 F. The steels were heat treated to a bainitic structure and tempered for a minimum of 24 h at 1100 F. The composition, including antimony, arsenic, and tin, along with the mechanical properties of each steel in the specimen, are included in the data. The temper embrittlement was measured as a shift in transition temperature as determined by the conventional V-notch Charpy test. The materials studied include both silicon and vacuumcarbon deoxidized steels, and a difference in rate of isothermal embrittlement is noted. The silicon deoxidized steels show a greater rate of embrittlement, and total embrittlement, at 850 F than at 750 F; vacuum-carbon deoxidized heats do not show this difference. The maximum amount of embrittlement during isothermal exposure is shown to decrease with increasing exposure temperature, with embrittlement very small at 1050 F exposures. The reversibility of the embrittling reaction is demonstrated using material that had been isothermally embrittled for 4800 h at 850 F. Removal of most of the embrittlement was accomplished by 20-h exposure at 1050 F, and embrittlement was measured again after isothermal exposure for 100, 1000, and 2800 h at 850 F. The reaction is shown to be fully reversible, but little insight into the mechanism is gained. In addition, it is demonstrated that exposures of both embrittlement and nonembrittled steels at higher temperatures in the embrittling temperature range (1000 to 1100 F) cause the transition temperature to shift to a common value; the nonembrittled steel demonstrating rapid embrittlement, while the embrittled steel de-embrittles to the same transition temperature which is apparently characteristic of the final exposure temperature. It can be shown that temper embrittlement, measured as an increase in transition temperature by the conventional V-notch Charpy test, can be suppressed by avoiding long exposure in the temperature range 750 to 1000 F during heat treatment. However, if long exposures in the 750 to 1000 F temperature range do occur either during heat treatment, fabrication, or subsequent service, it appears that temper embrittlement of present commercial Ni-Cr-Mo-V steels is unavoidable.