Комментарии:EAS 91, KW 308, SKE 220 (Doerfer's skepticism in TMN 3, 255 is hardly justified). The variation of *-g- and *-k- in Turkic and Mongolian is suspicious (cf. also the Nanai form pointing rather to *-k-), perhaps indicating that we are dealing with a merger of two different roots - cf. also the variety of reflexes in Korean; cf. also *sak`o and *zi̯úko.