does the triality symmetry (or something...) of d4 extend in some interesting way to f4, along the inclusion of d4 as the long root sub-system of f4??

i also vaguely wonder whether this might relate to some semi-mysterious comment of mckay's that i never quite understood...