Comments:АПиПЯЯ 281, Дыбо 11. Other comparisons of the above roots (Poppe 61, KW 335, VEWT 425a, SKE 233,240) are quite unreliable. The Kor. form sik- may go back to *sig-Vk- (cf. Turk. *sogɨk); a similar prototype can explain Manchu šaxurun < *šig-ak-urun. It is also interesting to note WMong. (L 658) saɣsurɣa 'hoar-frost' ( < *sug-sarga or *suɣa-surga) {and PT *sạgu-k}.