The probability table method (PTM), used in unresolved resonance region calculations, assumes that cross sections are not energy correlated. Strong cross-section energy auto-correlations are noted for some heavy nuclides that could affect the use of the PTM in the unresolved resonance region or its extension to the resolved resonance region. Uranium-238 has strong cross-section auto-correlations and is considered a severe test material for the PTM. Monte Carlo calculations of capture rates in 238U at 500, 1000, and 2000 eV do not show differences between the PTM and exact methods within an ∼1% calculational uncertainty. These results show that strong auto-correlations do not interfere with the use of the PTM in the resolved and unresolved resonance regions.