No competent person to-day believes that the Oxus ever entered the Caspian bodily in historical times. The dominant modern theory about the Oxus in the Greek period, as put forward by Professor A. Herrmann, is that the river itself entered the Aral, as to-day, but that before reaching the Aral it threw off a branch into the huge Sary Kamish depression south-west of the Aral, and that that branch issued southward from Sary Kamish, flowed down the Uzboi channel, and entered the Caspian at Balkan Bay, admittedly the only point where a lost river could enter the Caspian. He relied on two things: (1) a study by W. (V. V.) Barthold of Arab evidence of the thirteenth to the sixteenth centuries, and (2) an article in German by W. Obrutschew, who explored the Uzboi and published his results in Russian in 1890, of which book his German article is a summary.

Herrmann has envisaged a regular lost river, perennial and large enough to carry shipping, not an occasional spill-way, though Obrutschew called it an overflow channel for Sary Kamish when it got full; but both were rather obsessed by the belief that they had got to explain the northern or Oxo- Caspian trade-route from India, which never existed (App. 14). It was unfortunate therefore that Obrutschew discovered two (dry) waterfalls on his line of route, which would necessitate unloading and reloading vessels twice, a point already rubbed in.