Balto-Slavic C-stems / long vowel endings

Just a thought, before I forget.

Could Slavic nogá "foot, leg" and no``gUtI "nail" be split forms of a
common paradigm?

We could have:
N. *h3nógh(W)-o:t-s,
A. *h3nógh(W)-ut-m.

[Perhaps an oblique featuring /s/, like in *méh1no:ts, *m(e)h1nésos and
the ptc. in *-wo:ts, *-usos]. The N. regularly gives Balto-Slavic
*nagó:ts[*] > Slavic nogá:, while the A. yields BS *nágutiN > Slavic
no``gUtI.

[*] Which reminds me: I believe there was a very early Balto-Slavic
soundlaw whereby a long vowel (not contracted, not laryngeal) in the
ending attracted the stress. That explains why PD and AD nouns
consistently have an oxytone Nsg. (classical example *h2ák^mo:n >
akmuõ), and why the i- and u-stem (mobile) Lsg. is oxytone in (Early)
Slavic (from PIE *-e:i and *-e:u/*-o:u).

The present tense 1sg. ending *-o: does not attract the stress (Slavic
be``roN), which would, alas, be yet another argument against *-o:(u)
and in favour of *-oh3.

This would also resolve the discussion we had some time ago
about "mother": whether PIE had *máh2tV:r or *mah2té:r, Balto-Slavic
surely had *maHté:r (c.q. *maHté~), which subsequently became
*má:te:r/*má:te~ after Hirt's law.