From: tgpedersen Message: 40347 Date: 2005-09-23
> Or rather, both *k and *h2 together colour the > intervening *e to /A/. Still, the phoneme must be > *e here. >