After considering this couplet at intervals for over thirty years, I feel continually more and more strongly that Elmsley's explanation is the right one, i.e. that the construction is o ύ µή πOTε Tà σà έKφήνω kakà τνa ετπω Tà έµà έ πη, and that the second µή is only an emphatic repetition of the first.