Until now, there has been a discrepancy between the computed and the measured values of the 238U effective capture integral. The former has always been greater than the latter. For this reason, the reactor physicists have used an adjustment of the computed value. At the present time, the accuracy of the cross-section knowledge has increased, and the reactor computation codes are almost exact. Such an adjustment, therefore, is no longer justified. Recently, several new measurements of the resonance parameters were carried out and the use of a multilevel formalism was suggested to compute the 238U cross sections. This paper shows that the simultaneous use of recent parameters and the Reich-Moore formalism explain the discrepancy. For thermal neutron reactors, and depending on the neutron spectrum hardness, between one-half and two-thirds of this discrepancy is explained by the neutron data and the remainder by the multilevel formalism. This last effect is not negligible. We have done similar studies for 232Th, but in this latter case the multilevel effect was found to be much smaller than for the 238U and can be neglected in most applications.