Indeed, the next example seems to deny it to both. This example occurs, too, not far away from the scene of the first, in the twin bays of the Sabaeus Sinus, from which depart, _mutatis mutandis_, the double Hiddekel and the two Gihon. These twin gulfs bear so little imprint of being other than natural formations, that they have been universally and very likely quite rightly taken for such ever since Dawes discovered them in 1859, long before things like canals were dreamed of. It is strange that when the Hiddekel and the Gihon were found by me to be double in 1897, with a branch of both leading from each bay, the connection between the sceptically scouted doubles and the thoroughly believed-in bays should have been apparent. For to link a ghost to materiality, if it does not discredit the materiality, serves to substantialize the ghost. Furthermore, it shows that in this case neither the one double nor the other can have had its width engineered on any preconceived scale, unless the twin bays be themselves so accounted for. So that it seems useless to seek for cryptic standards in the canals or to think to find them a measure of value from the fact of their being a medium of exchange. [Illustration: The Sabaeus Sinus, embouchure for the double Hiddekel and Gihon.] A third instance of the same thing in the case of the Ganges and the Jamuna was proved at the last opposition after having long been suspected without my being able to make sure of it. These instances, taken in connection with the wide range of values in the widths presented by different canals, serve to show that the distance between the twin lines is an individual characteristic of the particular canal, and further to point to its cause, in some cases certainly and possibly in all, as topographical. The duplicate line makes a convenience of a neighbor, and suits its distance from its fellow to friendly feasibility. To cut a ‘canal’ to conform to the country seems logical if not obligatory, and quite in keeping with the nomenclature of the subject; but here the starting-point appears to be the only thing considered—the canal once safely launched being left to shift, or rather not shift, for itself. IV Topography thus introduced to our notice for its effect on the breadth of the doubles proves upon inspection to be of extended application to the whole subject. Examined for position these canals turn out to have something to say for themselves bearing on the question of their origin and office.