"Now in Avestan ‘lake/pool’ is vairi-. In fact Avestan has no verb or nouns cognate to the Sanskrit ones √sṛ >sar-. The only cognation is harah- in the name of the river harahvaiti – nothing else. But S √sṛ >sar- is a perfectly PIE morpheme appearing in Tocharian B sal-ate, Gk hiallō/hallomai and Latin salire all implying ‘moving, jumping’. Avestan somehow lost this root and its derivatives. So, how did the Iranians manage to concoct this name Harahvaiti, that sounds so exactly like a transliteration of the Vedic Sarasvatī , when they had no words from √sṛ > sar-? Adherents of the AIT offer no rational answers. (For full discussion see Kazanas 2004b, Prabhakar 1994/1995.) There is only one possible explanation. The Iranians, having lived in Saptasindhu moved to Iran (retaining the memory of the place as Haptahәndu); on meeting an amenable river there they gave it the name of the river they had formerly known – Sarasvati > Harahvaiti. The AIT can in no way, except by violating rationality, explain the two Avestan names Haptahәndu and Harahvaiti. This, if nothing else, should have alerted the AIT adherents to the possibility that there is something very seriously wrong in their migrationist scenario. Moreover, it is part of the general linguistic theory that the Avestan h derives from PIE s: so, it is again extremely difficult to see how the IAs who moved further southeast, retained the original s (in saras and saptasindhu) while the Iranians changed it to h."