In the first place, the canals have shown themselves, as they showed to Schiaparelli, to be seasonal phenomena. This negatives afresh the possibility of their being cracks. But furthermore, their seasonal behavior turns out to follow a law quite different from what we know on earth and betokens that they are indebted to the melting of the polar cap for their annual growth, even more directly than to the sun, and that vegetation is the only thing that satisfactorily accounts for their conduct. But again this is not all. Their time of quickening proceeds with singular uniformity down the disk, not only to, but _across the equator_. Now, this last fact has peculiar significance. So large are the planetary masses that no substance can resist the strains due to the cosmic forces acting on them to change their shape till it becomes one of stable equilibrium. Thus a body of planetary size, if unrotating, becomes a sphere except for solar tidal deformation; if rotating, it takes on a spheroidal form exactly expressive, as far as observation goes, of the so-called centrifugal force at work. Mars presents such a figure, being flattened out to correspond to its axial rotation. Its surface, therefore, is in fluid equilibrium, or, in other words, a particle of liquid at any point of its surface at the present time would stay where it was, devoid of inclination to move elsewhere. Now, the water which quickens the verdure of the canals moves from the neighborhood of the pole down to the equator as the season advances. This it does, then, irrespective of gravity. No natural force propels it, and the inference is forthright and inevitable that it is artificially helped to its end. There seems to be no escape from this deduction. Water flows only downhill, and there is no such thing as downhill on a surface already in fluid equilibrium. A few canals might presumably be so situated that their flow could, by inequality of terrane, lie equatorward, but not all. As we see on the earth, rivers flow impartially to all points of the compass, dependent only upon unevenness of the local surface conditions. Now, it is not in particular but by general consent that the canal system of Mars develops from pole to equator. From the respective times at which the minima take place, it appears that the canal-quickening occupies fifty-two days, as evidenced by the successive vegetal darkenings to descend from latitude 72° north to latitude 0°, a journey of 2650 miles. This gives for the water a speed of fifty-one miles a day, or 2.1 miles an hour. The rate of progression is remarkably uniform; and this abets the deduction as to assisted transference. The simple fact that it is carried from near the pole to the equator is sufficiently telltale of extrinsic aid, but the uniformity of the action increases its significance.