Try to answer all questions here.

Latest trend is for more of a squall line and less discrete cells. If that plays out then tornado threat will be slightly reduced.

Memphis will definitely be in the middle of things. Timing for Memphis looks to be 6:00-9:00. I hate that wide range but with the current disagreement on the models it's hard to get more specific. If it is mainly linear the later time will be more accurate. If we still get a discrete mode then the earlier time.

The Starkville area looks to be 9:00-midnight with the same caveat as Memphis.