>I think *jork(´)-o- 'roe deer' is a loanword from the NEC root *GHwo:r[ts']o 'deer, game' I've already mentioned. This "centumization" is found in other Vasco-Caucasian loanwords already described by Starostin in an old article.
> The little roe and the big red deer are consistently distinguished by
> those familiar with them (not only by Robbie Burns, "a-chasin' the wild
> deer, and followin' the roe"). <hart> and <roe> both have PGmc.
> protoforms, *xeruta/u-z and *raix-an- (f. *raix-o:n-), respectively. The
> former looks descriptive (related to the PIE 'horn' word-family), the
> latter is unclear, possibly a metathetic variant of the original term
> (see below). If one can speak of PIE terms for the two species, the most
> likely candidates (with the largest cognate sets) are *h1elh1-on- (vel
> sim.) 'red deer (stag)' (*h1elh1-n-ih2- for the hind) and *jork(^)-
> 'roe-deer' (which may have given pre-Gmc. *roik- by metathesis).
> Branch-specific replacements of either or both by new terms are commonplace.