A remark on the precocity of scaling in deep inelastic electron proton scattering