On the activation energy of the dc conductivity of organic liquids