Uranium fueled thermal-reactor lattice benchmarks, as well as some other experimental assemblies, were analyzed with ENDF/B-IV data and calculational techniques based on integral transport and SN methods together with Monte Carlo calculations in the resolved resonance region. Only a relatively small overprediction of the 238U resonance events—∼3 to 4%—remains when the present version of the ENDF/B data is used. It accounts for a considerable part of the ∼1% underprediction of criticality. Uranium-235 resonance absorption was found to be influenced noticeably by shielding due to the 238U resonances. For both 235U epithermal and 238U fast fissions, the agreement between calculation and experiment, although relatively good, showed greater fluctuations than in the case of the 238U capture. Calculated temperature variations of material buckling with temperature were greater than in the measurements, especially near room temperature, but the discrepancies were smaller in the critical than in the exponential assemblies.