Reconstructed root form for `vapor', with an uncertain assumption, is
*kwep- `to smoke, cook, move violently, be agitated emotionally'
(Pokorny kewHp- 596, Watkins kwe:p- 45) and according to "An
Etymological Dictionary of Proto-Indo-European Language', revised
version of "Indogermanisches EtymologischesWörterbuch" by Pokorny is
*we:p-/wH1p- (*swekW-) `to blow, to soar': Here belongs Skt causative
vapayati `make blow': Latin vapor (< *vapos) `vapor, heat' and Alb
vapë `summer heat'. Mallory-Adams reconstruct root *wapo:s based on
Latin vapor and Skt vaspa `vapor, steam, tears'.
I think that also Alb ufëm `big killing heat, which takes a breath',
as well as duhmë `oppressive atmosphere: (smell) stink; hot weather,
sultry heat, mugginess' are related here, but from different root
forms. If Alb vapë `summer heat' is derived from *wH1p-eH2, I think
that Alb ufëm is derived from suffixed zero-grade form *up-om with p-
> -f- in intervocalic position and duhmë is a prefixed form of ufma,
cf. also duf `breath, vapor, atmosphere; big anger, rage: buff',
restoring so the primary form of root *dheu-. So, ultimately I think
that if we agree that for Proto-Illyrian was characteristic the
dropping d(a)-, d(e)-, then pre-form should be *dheu-/dheuH-, a base
of derivatives meaning `to rise in a cloud, breathe', like *dhu:-mo-
`smoke': Latin fumus, Greek thu:mos, Alb tym, Sl dim etc. So, supposed
form *wap- or *wH2p- seems to be just an extended form of *(dhe)-uH2-p-
Konushevci