A survey is made of existing integral experiments for U233 systems and thorium-uranium based fuel systems. The aim is to understand to what extent they give a consistent test of ENDF/B-IV nuclear data. A principal result is that ENDF/B-IV leads to an underprediction of neutron leakage. Results from testing alternate thorium data sets are presented. For one evaluation due to Leonard, the results depict a possible growing discrepancy between measured integral parameters such as rho/sup 02/ and I/sup 232/ and the differential data, which underpredicts these parameters. Sensitivities to other nuclear data components, notably the fission neutron spectrum, were determined. A new harder U233 spectrum significantly reduces a bias trend in K/sub eff/ vs leakage.