Conductivity measurements in various polycrystalline ferrites have shown that the temperature dependence of the conductivity has to be described by two different activation energies, one which is attributed to the bulk material and the other which is due to the poorly conducting layers introduced by Koops. Activation energies as derived from the results of magnetic relaxation measurements of the initial permeability for samples of the BaFe12−2&dgr;3+Fe2&dgr;2+O19−&dgr;system are compared with those determined from the conductivity measurements. It appears that the activation energy as derived from the dc conductivity should not always be related to those determined from the magnetic relaxation. Both the high‐frequency conductivity and the magnetic relaxation losses are related to an electron transfer between Fe2+and Fe3+ions. The activation energies depend on the Fe2+content of the sample.