If you take speech utterance and find pdf of F0 and pdf of log F0. You will see log(F0)'s pdf fits more towards Gaussian. This might be one of the possible reason. Another possible reason is that sensation to sound is logarithmic.
-- Thanks & Regards, Nirmesh J. Shah M.Tech ( ICT)