In 1905 much more was made out about it, training in the subject and increased proximity of the planet contributing to the result. It now became clear to me that the canal did develop from the western side; for the western edge made a dark line of definite boundary from which shading proceeded to the eastern side, where it faded almost imperceptibly off with no defined line to mark its limit. That this shading gradually darkened was evident, but that when it could be seen at all it extended to the extreme limit of the eventual double, restricted the character if not the fact of an actual widening. At this opposition, too, the canal passed through its period of minimum visibility and was then seen, whenever it could be caught, as a confused swath of full width. In the case of this canal, then, a widening in the sense of a bodily separation of two lines seems inadmissible. On the other hand, the gradual darkening of the swath, and especially the advance of the darkening from the western side, points to an interesting process there taking place. [Illustration: Peculiar development of the Ganges.] At the opposite end of the series stands the Djihoun. As the Ganges is the widest of the instantly impressive doubles, so the Djihoun is the narrowest the eye has so far been able to make out. Only two fifths of the width of the Ganges pair, this slender double is very nearly at the limit of resolvability. So well proportioned are its lines to the space between them, however, that in ease of recognition it surpasses many wider pairs. In form, too, it is distinctive, turning by a graceful curve the trend of the Margaritifer Sinus into the Lucus Ismenius. With its fundamental branch—the northern of the two—it joins what is evidently the main line of the Protonilus—also the northern one—to the Margaritifer Sinus’s tip. [Illustration: Djihoun, the narrowest double.] It differs from the Ganges in some other important particulars besides width. In its case no band of shading distinguishes it at any time. It has always been two lines whenever it has been seen other than as a single penciling; the only confusion about it being evidently our own atmosphere’s affair. These two lines, furthermore, have showed, within the errors of observation, always the same distance apart. So that not only no change of intercommunication between the lines but no change in their places apparently occurs. Between these extremes in width, two hundred miles more or less for the Ganges and seventy-five miles for the Djihoun, the distance parting the pairs of most of the double canals lies. From 3° to 3°.2 on the planet may be taken as that of the average; the degrees denoting latitudinal ones on the surface of Mars, the length of which is equal to thirty-seven of our English statute miles.