When the attractions of Crocodilopolis and its suburbs became more appreciated and the population increased, the want of a larger area of cultivable land would be felt. There would also be another inconvenience, besides scarcity of arable land, felt by the dwellers in Crocodilopolis, arising from the yearly fall of the water surface. At high water, when the lake was filled up to R.L. 22·50, embarkation and disembarkation from boats might take place at Crocodilopolis itself, but, as the waters of the lake were allowed to flow back to the Nile, and the water level fell to R.L. 20·00 or 19·50, there would be laid bare a muddy margin of 2 kilometres breadth between the city and the water, which could with difficulty be crossed, and if crossed, the depth of water along the edge of the lake would be found too shallow to allow boats to get close to the land. One or both of these wants probably was the cause that led to the construction of the bank from the high land, east of Edwah to Biahmu, and thence, it appears probable, to Medineh. (See Plate XX.) The bank from Edwah to Biahmu runs generally along contour R.L. 17·50, and therefore would have been formed in water, probably with material transported from the high lands on the east and south-east of Edwah. This may account for the material of which the bank is formed being different from the land on either side of it, and for the absence of any trace of a borrow pit from which the bank was made. Such a bank, connected with the high land east of Edwah, running along contour R.L. 17·50 and joined from Biahmu to the high land at Crocodilopolis, would have enclosed an area, from which the lake water would have been excluded, the other two sides of the enclosure being formed by the natural ridge at the end of which Crocodilopolis was built, and by the high land connecting this ridge near Hawârat-el-Maqta with the commencement of the artificial bank at Edwah. From the levels of the rock underlying the Nile deposit at Hawârah it seems probable that the entering waters flowed in greatest volume past the Hawârah pyramid, separating the reclaimed tract from the desert on which the pyramid stands. Possibly this was the only channel by which the waters were admitted to the lake, and across which the regulator was built in the immediate neighbourhood of the Labyrinth and pyramid. The present course of the Bahr Yûsuf beyond Hawârah may have been closed and the Medineh ridge connected with the high desert on the left of the Bahr Yûsuf, near the modern head of the Gharaq canal. At present it is so connected, and the connection is only broken by artificial canals cut through it. [Illustration: PLATE XX. DIAGRAM SHOWING THE LIMITS OF LAKE MŒRIS ACCORDING TO THE THEORY FAVOURED IN THIS PAPER. _Note._—The contours along the N.W. side of the lake are not accurately shown as this side has not been surveyed.