On the spectrum of the secondary auroral electrons