Brook Conner teithant:
> A somewhat less orthodox mode might allow tehtar to be drawn
> *below* a tengwa as well as above. If you do this, then a pair of
> vowels bracket one tengwa.
That's attested in the Edwin Lodham's Manuscripts, DTS 50 and 51, and
it's interesting that they don't show the same tengwar-tehtar order.
DTS 50 has: 1st tehta above, 2nd tengwa, 3rd tehta below; DTS 51 has:
1st tengwa, 2nd tehta above, 3rd tehta below. You may call it "less
orthodox", but only because that's a rather unusual mode, not because
it were not attested.