Right. Etruscan may have a connection with Rhaetic in the Alps, but that’s not quite established, and the only sure connection is Lemnian, as you mention. I did a little amateur’s analysis on the Lemnian stele years ago, with Bonfante’s Etruscan grammar in hand.
There is really not that much difference at all between the two idioms...they look to me like two slightly different dialects. From what I could see, they didn’t look like languages that had been separated for thousands of years—which sort of puts a crimp in the idea that they represent a pre-Indo-European substrate. Impressionistically, I’d say that the separation between them looks to be only on the order of 500 years or so.
And your point about the languages retreating to the hills and inaccessible areas is certainly well-taken. That’s exactly what happened to Oscan, as the Samnites were deprived of the coasts by Greeks and later Romans.
If the Etruscan parent population had died out by the time of Dionysius, that would explain why he didn’t see any connection. I’m pretty sure Lemnos already lost their language several hundred years prior to his time.
Herodotus in 6.136-140 tells how Miltiades (later hero of the battle of Marathon) conquered Lemnos for Athens. His account implies that some of the Lemnians were forced to leave the island, but maybe not all of them.