On the Arabian Frontier of Egypt. 41 



dent that this entire tract must have sunk below its former 

 level since those cities were habitable. 



Owing to this depression having cut across the Pelusiac 

 arm, its waters could not follow the old channel up to its 

 month, had it remained open. During the flood season, the 

 surplus water still flows along the extinct arm, as far as 

 Fakkoos (^Phacusa), beyond which it is conveyed under the 

 sand in a direction considerably to the south of the former 

 channel, dividing near Salahieh into two branches that ter- 

 minate in the marshy lake Ballah, which forms part of Lake 

 Menzaleh at that time of the year. 



The waters of the Etham branch would also run no further 

 than this lake, were its course to be now restored. It is there 

 that they were lost, in the accidental irruption of 1800. Mr 

 Devilliers, in his journal, says, " Un sheik nous dit: ' Has el 

 Moyeh el-Ballah a vu I'eau du Nil cette annee.' " (vol. xviii., 

 p. 380.) But before any additional depression of that already 

 very low tract had taken place, the centre of the hollow 

 formed by the continuation of the ancient strait northwards, 

 and now occupied by a line of salt marshes and by the above 

 mentioned Lake Ballah, must have marked the bed of the river, 

 as far as Migdol, and it may then have passed round tlie 

 hill of Migdol, and run into a lake now dried up and filled 

 with sand, which is the Lake SerOonis of Herodotus. During 

 the inundations, a line of sweet water under the sand, appa- 

 rently the continuation of a deep eastern inlet of Lake Ballah, 

 and along which there are several wells, seems to point out tlie 

 position of a former watei'-course extending in the direction 

 of the modern village of Katieh. The firmness of the soil 

 along this line, and abundant vegetation, seem to countenance 

 this supposition. Nevertheless, I must acknowledge that 

 this course of the river beyond Migdol rests wholly on in- 

 ferences drawn from such physical indications as a map 

 esteemed a masterpiece of tojjograpliical accuracy in details, 

 affords — for such the great Frencli map of the Delta is ad- 

 mitted to be. The survey of the northern lialf of the Isthmus 

 did not follow the line of the lowest points. It cut across 

 two considerable swells of ground, at the base of which the 

 water courses lay. A special observation alone can deicr- 



