I have explained the first member in my article in KZ., 68, 218 f. (relationship to Arm. eres). The meaning of the second member is certainly near that of the first, since we evidently have to do with an old dvandva. I would suggest connection with the root of Olr. Mlr. bind “ melodious ’, Skt. bhand- “ worship, exhilarate, shine ”, IE. *bhend- “ ± sing ” (Walde-Pokorny, 2, 151). For the semantic development “ ± countenance ” from “ ± sing ”, cf. Skt. vadana- “ mouth, face ” and vad- “ speak, sound ”, and also Toch. A. sunk-, B. sähkw- “ muzzle ”, related to Grk. όμΦ “ voice ”, Goth, siggwan, OHG., OS. singan “ sing ” (cf. my Lexique etymologique des dialectes tokhariens, 127, and my Morphologie comparee du tokharien, 45).