Kottish:fugai, phugai, *fukai, pl. fugajaŋ, *fukajaŋ; Ass. (М., Кл.) pugáj
Arin:ṕhugáj (М., Кл.); isbugej (Лоск.)
Comments:ССЕ 253. A certain derivative is Kott. fugaʔiše, *fukajaše, Ar. pogejši (Стр.), ṕhúgajši (М., Сл., Кл.) "sable". Werner 1, 332 <*phuk-at / *phuk-aj>. The root contains a frequent body part suffix *-Vʒ - which, despite Werner ibid., has nothing to do with Ket aʔt 'bone'; nothing like *huk 'flaumig, wollig' is ever attested in any Yenisseian language. Werner's statement that "dem ket., jug. aʔt 'Knochen' entspricht in Komposita das kot. -ai" is inexplicable phonetically and plainly wrong: in a few cases of attested compounds of this type Kottish has not -ai, but -ar: cf. *χɔb-aʔd 'back(bone)' > Ket qɔvǝt, Yug xɔfat, Kott. hapar, *kas-aʔd 'foot sole' > Ket. kassat, Kott. hačar - quite consistently with the Kott. reflex of *ʔaʔd 'bone' which is Kott. ar-aŋ, ar-ša q.v.
Comments:ССЕ 253-254. The Ket and Kott. languages point to PY *puχV (Werner 1, 347 <*phɨ>, not taking into account the Kott. -g-). The Arin matches are difficult in two respects: a) the word means both "nephew" and "brother/sister" - the latter must be a secondary semantic expansion of "brother's son" to "brother" in general, having replaced the original PY *b[i]s "brother / sister"; b) the -m- in Arin must reflect an original nasal, preserved after the prefixes a-, ba- and replaced by *p- elsewhere in Yenisseian. Phonetically the case is similar to the root for 'brain' *nɔʔŋ > *dɔʔŋ, which is preserved as *nɔʔŋ in the compound *goʔ-nɔʔŋ, but has regularly changed to *d- in independent position (Ket, Yug dɔʔŋ). Werner's objection: "die letztere Rekonstruktion [*m[u]χV] bleibt fraglich, weil sie sich auf ar. bamagal 'Bruder' stürzt, wo das anlautende b- eigentlich das Possessivpräfix der 1 P.Sg. ('mein') darstellt" - is rather strange: of course b- is a 1 p. prefix (as in numerous other cases), but the reconstruction is based not on it, but on the following -m- - which cannot be a prefix of any kind. Elsewhere (1, 32) Werner attempts a different approach to the Arin form, analyzing it as am "mother" + agel (?), cf. Pump. akil "older brother, sister". Pumpokol indeed has such a word (see under *ʒVl 'child'), but Arin has it too, namely Arin akel 'son', akelä 'daughter' (see ibid.). Whereas "brother" is of course "mother's son", "nephew" definitely is not: the compound am + akel 'mother's son' could have helped Arin amagel (amakel) to shift its meaning to include both "nephew" and "brother", but it cannot explain the meaning "nephew" in any reasonable way.
Comments:ССЕ 255. The etymology is not quite reliable. The status of -t- in Yug is unclear (in Ket this consonant could have been dropped within a cluster): if it is not a part of the root, the Ket-Yug and Kott. forms can be hardly related, and one could only reconstruct two different roots: *pVn- for Ket-Yug and *put for Kottish. The latter solution is perhaps preferable, since the correspondence Ket-Yug -nt- : Kott. -t- is irregular (the cluster *-nt- should be simplified in Ket-Yug, but preserved in Kottish). The stem may be connected with *pǝnVŋ 'sand, ashes' ("grind" = "reduce to sand"), although derivational details remain unclear. Werner 1, 315-316 also prefers to reconstruct *phin- and leave Kott. -fut aside.