of the snow, these stars shone for a few moments and then slowly disappeared. The seeing at the time was very good. It is at once evident what the other-world apparitions were,--not the fabled signal-lights of Martian folk, but the glint of ice-slopes flashing for a moment earthward as the rotation of the planet turned the slope to the proper angle; just as, in sailing by some glass-windowed house near set of sun, you shall for a moment or two catch a dazzling glint of glory from its panes, which then vanishes as it came. But though no intelligence lay behind the action of these lights, they were none the less startling for being Nature's own flash-lights across one hundred millions of miles of space. It had taken them nine minutes to make the journey; nine minutes before they reached Earth they had ceased to be on Mars, and, after their travel of one hundred millions of miles, found to note them but one watcher, alone on a hill-top with the dawn.

Calculation showed the position of the star-points to be in longitude 280 degrees and 290 degrees and in latitude 76 degrees south. At this place on the planet, then, there was a range of slopes sufficiently tilted to reflect the Sun from their ice-clad sides. On comparing its position with Green's map of his observations upon the cap at Madeira in 1877, it appeared that this was the identical position of the spot where he had seen star-points then, and where Mitchell had seen them in 1846, to whom they had suggested the same conclusion. Green christened them the "Mitchell Mountains." At the times both these observers saw them, they were detached from the rest of the cap. At the time of this observation in June, they were still in the midst of the cap. We shall see that they eventually became islands, just as Green saw them, and that the observation in June marked an earlier stage in their history.

On June 10 Mr. Douglass detected a second rift in the cap backing the range of slopes. And on June 13 I noticed that behind the bright points the snow fell off shaded to this rift. Meanwhile a third rift had been made out by him, running from longitude 170 degrees to longitude 90 degrees,--very nearly, therefore, at right angles to the first rift and debouching into it. Bright points continued to be seen at various points to the westward round the cap. They are marked by crosses on the chart. Throughout these days, the cap was wont to appear shaded upon the terminator side, as one might expect of a snow or ice slope. During June, also, the contour of the cap was apparently elliptical. But on June 25 Professor W. H. Pickering noted, for the first time, that it no longer looked so. The melting had resulted in making its asymmetry perceptible.

On July 1 our Martian polar expedition disclosed what used to be the supreme quest of earthly expeditions,-- that dream of arctic explorers, an open polar sea. On that day Professor Pickering perceived, in the midst of the cap, in longitude 260 degrees and latitude 80 degrees, a sheet of water about 250 miles long by 150 broad. It was in fact the spreading of the first rift about midway across the cap, and lay not far from the geographical pole of the planet, though not, it is to be noticed, near the pole of cold for it lay on the further side of the geographical pole from it. There is a touch of the irony of fate in this detection of an open polar sea on Mars before explorers have succeeded in doing so on the Earth.

In addition to these rifts and other irregularities of melting, small detached bits of the cap showed from time to time, one being seen by Professor Pickering on July 9 in longitude 284 degrees, and another by him on July 23 in about longitude 160 degrees.

Meanwhile the cap had been steadily decreasing in size, its progressive diminutions being shown on the map in the successive contour lines. The polar sea faithfully followed it in its shrinkage, even the bays keeping their longitudes unchanged. But, whereas early in June the bay at longitude 270 degrees had been blue, it now appeared brown; of that mud-color land does from which the water has recently been drained off.

After various vicissitudes, too numerous to mention in detail, on August 6 a separate patch of snow showed very conspicuous, to the left of the main body. The smaller detachment lay in longitude 290 degrees, and in latitude 75-78 degrees. Now, on turning to the record of the star-points that had appeared two months before, it will be seen that this was their position. Here, then, was proof of the identity of the star-points seen in June with the islands recorded by Mitchell and Green. The detached patch was in