In Ebers' delightful book on Egypt space is given to the description of the much more modern one located at Rôda. The nilometer, or "mikyās," on the island of Rôda, now visible, is stated to have replaced one which was brought thither from Memphis at some unrecorded date. Makrīzī in 1417, according to Ebers, saw the remains of the older nilometer. The present mikyās is within a covered vault or chamber, the roof being supported on simple wooden pillars. In a quadrangular tank in communication with the river by a canal is an octagon pillar on which the Arabic measurements are inscribed. These consist of the pic (variously called ell or cubit) = 0·54 metre, which is divided into twenty-four kirats. In consequence of the rise of the river bed in relatively recent times, the nilometer is submerged at high Nile to a depth of two cubits. [Illustration: SCALE OF THE NILOMETER AT RÔDA.] The rise of the Nile can now be carefully studied, as gauges are distributed along the river. We have the Aswân gauge from 1869, the Armant gauge from 1887, the Suhag gauge from 1889, and the Asyût gauge from 1892. The distances of these gauges from Aswân are as follows:-- Kilometres. Aswân 0 Armant 200 Suhag 447 Asyût 550 Rôda 941 The Rôda gauge is not to be depended on, as the movements of the barrage regulation destroy its value as a record. The heights of these gauges above mean sea-level are as follows:-- Metres. Aswân 84·158 Armant 69·535 Suhag 56·00 Asyût 53·10 Rôda 13·14 [Illustration: THE ISLAND OF RÔDA.] Great vagueness arises in there being no very obvious distinction between the gauge readings reached in summer and that from which the rise is continuous. There are apparently rainfalls in the end of spring of sufficient power to raise the Nile visibly in summer, just as muddy rises have been seen in winter to pass down the valley, leaving a muddy mark on the rocks at Aswân and Manfalūt. Independently of the actual gauge-reading of the rise, there are facts about it which strike every beholder. At the commencement of the rise we have the _green water_. This occurs in June, but varies in date as much as the top of the flood varies. From the fact that modern observations show that the very beginning of the rise, and the first flush, second flush and final retirement vary, it seems evident that the ancient Egyptians could not have had any fixed zero-gauge or time for the real physical fact of the rise, but must have deduced from a series of observations either a mean period of commencement, or a mean arrival of the red water, or a mean rising up to a certain gauge.