Actually, it was *t-t-, not *t-r- in the protolanguage, as I have explained before. The -r- results from t-voicing in _some_ daughter languages.
 
----- Original Message -----
From: tgpedersen@...
To: nostratic@yahoogroups.com
Sent: Wednesday, April 11, 2001 6:45 PM
Subject: [nostratic] Re: ouralique et IE

Yes, them dry bones... There's the Sahara to contend with.
As for the "3" *t-r- in Niger-Congo, if you are thinking of that...
Piotr tried to reduce it to a Proto-Bantu thing, but obviously it's
bigger than that. But it is not omnipresent in Niger-Congo, it occurs
in some languages and not in others; and isn't that the situation
you'd see if that term were borrowed?