Dose measurements using Fricke and ionization methods were compared for 60Co gamma rays, 4-25-MV photons, and 10-25-MeV electrons. Fricke derived doses based on a constant yield (epsilon mG) were in good agreement with ionization derived doses based on the American Association of Physicists in Medicine Task Group 21 protocol and the National Research Council of Canada (NRC) ND calibration or the NRC proposed Nx. These measurements also confirmed the validity of the double-voltage technique in the collection efficiency correction, even for swept electron beams. Assuming the correctness of the ionization derived doses, the radiation yield appeared to be 1% higher and to increase with photon energy when irradiation vessels were made of Pyrex but not with polystyrene cells. These glass wall effects could be due to the scattering perturbation of electrons between inhomogeneous materials and, in particular for photon beams, due to the mismatch in mass energy absorption ratios and mass collision stopping power ratios between the Fricke dosimeter and the wall materials.