and far outshone the eastern half, being almost as brilliant as the polar cap. When he looked at it again the next night, September 25, the effect of the night before had vanished, the western half being now actually the darker of the two. So fugitive an effect suggests cloud, forming presumably over high ground, and subsequently dissipating; it also suggests a deposition of frost, melting on the next day. It is specially noteworthy that the canals inclosing the region, the Galaxias, the Boreas, and the Eunostos, were not in any way obscured by the bright apparition. On the contrary, Mr. Douglass found them perceptibly darker than they had been, an effect attributable perhaps to contrast.

Although not storm-clouds, it is possible that these appearances may have been due to thin cloud, capping high land. There are objections to this view, but as there are graver ones to any other it may stand provisionally, the more so that there are appearances not easily reconcilable with other cause. For example, a most singular phenomenon was seen by Mr Douglass on November 25, a bright detached projection, for which, from measurement, he deduced a height of thirty miles. This would seem to have been cloud, for the details of its changes in appearance seem quite incompatible with a mountainous character. With regard to its enormous height, it is not to be forgotten that a few years ago on the Earth phenomenal dust- clouds were observed as high as one hundred miles.

We now come to a highly interesting class of observations bearing upon the question of clouds,--Mr. Douglass's terminator observations. During the last opposition, seven hundred and thirty-six irregularities upon the terminator of the planet were detected at Flagstaff. They were seen by one or more of three observers, but chiefly by Mr. Douglass, who made a systematic scrutiny of the terminator for almost every degree of Martian longitude. Their full presentation would be both too tabular and too technical for this book. The paper embodying them will be found among the published annals of this observatory. I shall here give only certain deductions from it.

Of the 736 irregularities observed, 694 were not only recorded but measured. Of these 403 were depressions. It is singular, in view of their easy visibility, that they never should have been noticed before. Schroeter, indeed, saw three appearances of the sort,--on September 21, 1798, November 12, 1800, and December 18, 1802,--but all on the limb, not the terminator, which shows them not to have been of those here meant. Nevertheless they are not difficult to see, and anything but rare. When the phase is large enough, several may be seen every night.

The projections number 291. As their number shows, they are less common than the depressions, but they are even less of a feature of the surface than their number would indicate, for the depressions extend as a rule much further both in latitude and longitude.

Usually the depressions look like parings from the planet's rind, and almost always appear upon that part of the terminator where the dark regions are passing out of sight; commonly therefore, in the case of the southern hemisphere, they are met with between latitudes 30 degrees to 60 degrees south. Not so common is it for them to occur over a part of the planet which is bright. Furthermore, they appear to occur more or less continuously. This would not be the case were they real depressions.

As this may not at once be evident to the reader, and yet is easily made evident, we will consider the diagram on page 38. It will there be seen that an elevation like s or r--and the same reasoning applies mutatis mutandis to a depression--appears projected a relatively long way without or within the terminator, as compared with its actual length, owing to the angles under which it is respectively illuminated by the Sun and seen from the Earth. The relation between its height and its distance from the edge is that between the height of a hill and the shadow it casts at sunrise or sunset. What, therefore, is not high enough to be seen in profile on the limb, becomes vicariously visible on the terminator. But a hill could not continue long to appear as an elevation, as the rotation of the planet would carry it in due course from the position r to the position s, and there it would be forced to masquerade as a depression. The same, reversely, would happen to a valley. In order that a depression should appear continuously, there must be a belt of lower level along its circle, and this could not be made visible as in the former case by projection, since projection depends upon difference of level along the same surface contour, not as