Remarks on a non linear equation arising in population genetics