PAT *pǝza (cf. also Bzyb. a-pǝ́za); PAK *cʷa-bʒá. The first part of the Adygh and Ub. compound, despite Shagirov 1977, 104-105, is connected neither with PWC *c:ʷǝ 'ox' (q.v.) nor with PWC *zʷa 'to plough' (q.v.), but reflects a separate PWC root (see under *Hrājc_ū). Again, pace Shagirov ibid., since the correspondence between Abkh. and Ad. here is quite obvious, it is impossible to connect the second part of the compound with PAK *bzǝ 'to cut' (q.v.). Finally, since Ub. has a palatal ź, the Ub. form (again despite Shagirov ibid.) can not be regarded as an Adygh loanword. The correspondence between PAK *-bʒa, PAT *pǝza and Ub. -bźa is quite regular.