Of what they look like, the following illustrations give a fair idea, only that instead of being more geometrically regular in the drawing than in reality the fact is the other way. Freehand draftsmanship at the telescope is incapable of rendering their ruled effect. No railway metals could be laid down with more precision. As to their size, the following figures derived from a typical double canal, the Phison, give some conception. This great artery of intercommunication between the Sabaeus Sinus and the Nilosyrtis is, roughly speaking, 2250 miles long; the distance between the centres of the two constituents is about 130 miles, and each line is perhaps 20 miles in breadth, when at its maximum strength. The pair follow, apparently, the arc of a great circle from the Portus Sigaeus on the Mare Icarium to the Pseboas Lucus in latitude 40° north. The Portus Sigaeus consists of two little nicks in the coastline, looking like the carets one makes in checking off items down a list, if the space between the down and up strokes were then filled in; the Pseboas Lucus, on the other hand, is a large round dot like a small ink spot. To these two differently appearing spots, the twin lines of the Phison behave differently. While each line leaves centrally its own caret of the Portus Sigaeus at the south, at the north each touches peripherally the Pseboas Lucus, on the east and west sides respectively, the two thus just holding the Lucus between them. In position the lines are invariable, though in visibility not. Sometimes only one is seen, sometimes both show faintly, and sometimes both are conspicuously strong. The delicacy of the observations by which this detail was established is second only to its importance. It destroys at a stroke all possibility of diplopic unreality, since were that the fact the Pseboas Lucus should be doubled, which it is not. At the same time it opens vistas into the true construction of the things themselves, at present more suggestive than satisfactory. [Illustration: Martian doubles (corroborating the above).] In the great circle character of its course the Phison is quite normal. The majority of the double canals pursue the like method, running straight over the surface from one point to another, the constituents remaining equidistant throughout. But such forthrightness of direction, though the rule, is not without exceptions. The Thoth-Nepenthes, for example, sweeps round in a seemingly continuous curve to the west-southwest from the Aquae Calidae to the Lucus Moeris like some mighty bow perpetually bent. Nevertheless its lines are no less careful for all their curving to keep their distance from one end to the other of their course. The quality of being paired rises superior to change of direction. II