A graphite flat cavity ionization chamber is used at the BIPM in France to determine the absorbed dose to graphite in a 60Co photon beam and thereby used to determine the product of the value of W/e, the average energy required to produce an ion pair in dry air, and the value of (LDelta/p)a(C), the mean restricted mass collision stopping-power ratio for graphite to air in a 60Co beam. The accuracy of the (W/e) (LDelta/p)a(C) value thus determined depends upon the accuracy of the perturbation correction factors adopted for this chamber. The perturbation effect of this chamber was accounted for by the replacement correction factor whose value was calculated by an analytical method and confirmed by an EGS4 Monte Carlo calculation. The purpose of this study is to investigate the validity of the analytical and the EGS4 calculations by using recently established methods and the EGSnrc Monte Carlo code, a much improved version of EGS4, to calculate the replacement correction factors for the graphite chamber. It is found that the replacement correction factors used for the BIPM chamber are not correct: the values used are smaller than they should be by about 1%. This leads to a 1% overestimation of the (W/e) (LDelta/p)a(C) value determined by using this chamber. This implies that 60Co air kerma standards that are directly proportional to this product need to be reduced by 1%. Based on the values of the replacement correction factors calculated in this study, and on the value of (LDelta/p)a(C) evaluated from ICRU Report No. 37 stopping power for graphite, the value of W/e determined by using the BIPM chamber should be 33.61 +/- 0.08 J/C. If a more recent value of mean excitation energy for graphite (86.8 eV) and grain density are used to evaluate the graphite stopping power, then the value obtained for W/e is 34.15 +/- 0.08 J/C.