"Rather, the Iranians left the region of the 7 rivers and held the name in their memory. Something very similar happens with the V river-name Sarasvat/ and Av Harahvaiti-. Avestan has no other cogn with harah- whereas S has sr > sarati/sisarti, sarana, saras, sarit, etc, etc and of course cognates are found in other IE branches: here again it is the Iranians that took with them the memory of the Indic river and gave it to a river in their new habitat... Moreover, Vedic retains the PIE s but this becomes h in Avestan. All this actual linguistic evidence and the conclusion it forces upon us has some archaeological/geographical support. G Gnoli, who is a normal AIT adherent and by no means an indigenist, showed very clearly that the early portions of the Avesta hardly know northern and western Iran and he analyses migrations there from south to north and east to west but not north- west down to south-east (1980). Thus while the conjectural Indo-Iranian movement south-eastward contains many anomalies, the Iranian movement from Saptasindhu north-westward accommodates all facts."