Last in time but not least in importance of the details of canal development to be detected is one that connects these strange features directly with the melting of the polar caps. The cartouches showed that such connection was to be inferred; the facts now to be recorded depict it by an identity of place between certain phenomena of the two subjects following one another in order of time.

On January 8, 1897, while scanning the planet, I was suddenly ware of a rift in the north polar cap. It ran a little to the west of south from where it started in at the cap’s edge and went clean through to the limb, the pole being then slightly tilted away from us. At the time it seemed to be the first rift ever seen in that cap; but on opening a little later Schiaparelli’s Memoria Quarta, which had just arrived, the first thing my eye fell on was a drawing of a rift in the north polar cap observed by him when the planet had held the like attitude toward the Earth thirteen years before. Reference to its longitude showed it to be the identical rift, seen again after all these years and the only one so far seen in the northern cap.

At the next opposition more rifts were detected, one in especial on December 27, running from Arethusa Lucus, then upon the edge of the cap, athwart the snow in a northwesterly direction.

In the forepart of the opposition of 1901, which in its Martian season corresponded to that in 1897, when the rift had been observed, many rifts were detected in the cap, and among them one traversing the cap north-northeasterly in longitude 136°.

So far the season when the cap had been observed was that when the rifts were in process of forming. The ground they and the snow-cap covered had not yet at any opposition been uncovered.

It was only when my observations began in the latter half of the opposition of 1901 that, the season on Mars having so far advanced, all snow in those latitudes had melted. Then appeared, however, the canal Hippalus, an arctic canal of some importance, lying on that part of the planet previously occupied by the polar cap. When later studying the observations on the rifts I remembered this canal, and turning to the drawing made some months before to compare the two critically, discovered that the canal occupied the precise position held earlier by the rift. One had said the rift had never vanished, but that the white surrounding it had simply turned to ochre. Here, then, was a striking coincidence of place, too exact to be the result of chance.

Impressed by the identity, I examined all the other rifts seen early in 1901, comparing them with the arctic canals seen later, to the finding of no less than five cases of the same coinciding positions.

The importance of the identification here made of an arctic canal with a previous rift in the polar cap has led me to make a list of the canals thus identified at this opposition.

If it be asked why these canals do not appear recorded at the March presentation as either the one phenomenon or the other, the answer is twofold. First, because they showed as shadings lost amidst a shaded mass; and, secondly, the observations at several oppositions indicate a great amount of haze over the region at that season of the Martian year.

We may now go back to the very first rift, that of 1897. The Martian season grew later with each succeeding opposition, and it so chanced, abetted by this fact, that the delaying snow was never seen covering that part of the planet again and so, of course, not the rift. The Martian summer in those high latitudes came on, and with it brought the great arctic canal, the Jaxartes, into conspicuousness. The canal in consequence had been observed for some time before it proclaimed itself the apotheosis of a rift and that of the first and most important rift of all. Comparison of position, however, entirely confirmed the conjecture and added another and the most striking of all to the list.

These six canals, on the whole the largest which run into the northern cap, have thus a dual character. Starting originally as rifts, they later come out unmistakably as canals. So that we may say in general that the two phenomena are different seasonal states of the same thing. This instantly explains the rifts, the origin of which we found of so difficult, not to say impossible, interpretation before in these pages, and incidentally it confirms what we deduced on other grounds as the character of the canals; to wit, strips of vegetation. For if the cap covered desert and fertility alike, it is precisely over the latter that it would first melt.

Vegetation has the property of melting snow. The metabolism of the plant, like that of the animal, though in a less degree, generates caloric. A living animal is warm, even the so-called cold-blooded ones, in some sort, and a growing plant is too. The chemic processes concerned give off heat, though in such small quantities that we are often not aware of it. While the plant lies dormant it stays cold, but the moment its sap begins to run under the rays of the spring sun it rises in temperature above its winter surroundings. All it needs to this awakening is sun and water, and both it gets in its place in the polar cap after the passing of the vernal equinox. The time, therefore, is suitable, for it is not till after that equinox is passed that any of the above phenomena occur. In consequence the snow about it melts and the plants themselves show as dark rifts splitting the cap.

This quite unexpected identity of two seemingly diverse phenomena, and the unsolicited support its only explanation lends to the general theory, is an instance of what is constantly occurring as observation of the planet is pushed farther and farther. Facts every little while arise which prove to fit into place in the scheme when neither the facts nor their fitness could have been foreseen.