extent silted up. In later Pleistocene times drainage down the Nile Valley appears to have become more pronounced, the barriers between the lakes were broken down and the river cut for itself a channel through the lacustrine beds, filling up the old lake basins. From this time onwards to early prehistoric times the bed of the river would appear to have steadily fallen, as it eroded its channel deeper and deeper. That this early Nile was a river of considerable size is evident from the amount of erosion it accomplished in the trough of the valley, whence the older lacustrine beds have been almost completely removed. Probably in the lower part of its course it swept against the base of the dividing ridge between the Nile Valley and the Fayûm depression. If, as we imagine, the river was at that time flowing some 20 metres higher than at present its currents would have met with little resistance from the loosely cemented terrace of gravel which formed the upper part of the dividing ridge. Once this was broken down the waters must have poured into the depression behind, until a lake of considerable size was formed. The sediment spread out over the floor in the shape of a fan, while at the same time deposits of sands and fine clays, blown and washed into the waters from the surrounding shores, were being slowly accumulated in the quieter and more remote parts of the lake. For some time subsequent to the first connection between the Fayûm depression and the Nile, the latter continued to fall in level owing to continued erosion along its course, possibly the Fayûm again became completely isolated for a time. Subsequently in the earliest historic times under changed conditions the river commenced to carry and lay down the modern alluvial deposit of “Nile mud”, and from this time to the present day its bed has gradually risen. This is shown by the high Nile flood-readings on the early gauges of the Nile Valley; the nilometer at Roda shows a difference of 1·22 metres in 1026 years, or an annual rise of 0·12 centimetre, which is equivalent to 12 metres in 10,000 years[90]. Whether there was ever complete disconnection between the Nile and the originally formed lake in Pleistocene times is uncertain, but even so it was probably only during a comparatively short period: in early historic times the rise of the Nile bed must have brought about a reconnection.