Comments: PAT *mǝgʷa; Ub. def. a-nǝ́ḳʷ. A complicated case of correspondences between AA and Ub.: we must assume a secondary deglottalisation in PAT (for reasons which are not yet clear) and a dissimilative change *m- > n- before a labiovelar in Ub. (we should note that roots of the type mVKʷV are indeed very rare in Ub.). Comparison of Ub. nǝḳʷǝ́ with PAK *ḳʷác̣ǝ 'intestines' (Шагиров 1, 217) is not persuasive.