Comments:KW 92. A Western isogloss. The Turkic and Mong. forms are no doubt related, but Ramstedt treats them both as derived from *deg- 'top'. Since we divide the traditional reconstruction of the latter root into *di̯òge and *tēga (q. v.), such a derivation seems less plausible. The Ud. form diŋe-, albeit isolated in TM, also supports deriving Turk. *jẹŋ- and Mong. dejil- from a separate root.