Comments:KW 352, Poppe 30, 46. Correspondences are quite regular, and there is hardly any need in assuming (despite Poppe 1966, 198, Doerfer MT 123) Tung. < Mong. However, Mong. may be borrowed from Turkic (see Щербак 1997, 148-150), in which case a reconstruction *síba is also possible.
