[en] In the NSRR experiments, it has been generally believed that an axial power distribution in a test fuel rod is almost flat due to the core length three times longer than a fuel stack region in a test fuel rod. However, the axial power distribution in a test rod is possible to change because of the effects of both control rod position and reflection of neutrons at edge regions of the fuel stack. Therefore, we measured the detailed axial distribution of γ-ray intensity in test fuel rods to characterize the power profile and to evaluate the effect of the power peaking on the fuel rod behavior. Based on the measurement of the γ-ray distribution in a fuel rod, the power was observed to be higher at the lower portion of a test fuel rod, and the effective power peaking was estimated to be 1.04-1.06 compared with the average in a fuel stack region. In the NSRR experiments, a fuel failure occurred usually at the lower portion when the fuel rod was subjected to the energy deposition of 260 cal/g UO₂ in axial average. On the other hand, reducing the peaking by insurting 5 % enriched UO₂ pellets at both edges of the fuel stack, a fuel failure did not occur even if the fuel rod was subjected to 260 cal/g UO₂ in axial average. This must indicate the strong effect of the peaking on the fuel rod failure behavior. As the relation was observed between fuel rod behavior and a local energy deposition in a fuel rod having a widely deviated power profile in axial direction, it must be better to evaluate the fuel rod failure behavior based on the local energy deposition. Researching the failed position among the test fuel rods subjected to around failure threshold energy of 260 cal/g UO₂, the local energy deposition was estimated to be around 274 cal/g UO₂ at failed position. (author)