It is not uninteresting to inquire in what the difference consists. The slight difference of tilt in the Martian axis would slightly extend the breadth of the tropical and the polar regions at the expense of the temperate ones, and thus accentuate the seasons, but the chief seasonal contrast between Mars and the Earth would come in consequence of the much greater eccentricity of Mars' orbit. For the more eccentric the ellipse, the greater the variation in the planet's velocity at different parts of it, inasmuch as the Sun pulls the planet toward himself with a force depending on his distance. The less this distance, the greater the angular velocity. But the angular velocity determines the length of the seasons upon a planet whose pole of rotation is tilted to the plane of its orbit, like the Earth or Mars. The greater the eccentricity of the ellipse, therefore, the greater the difference in the length of the seasons. In the case of the Earth the difference is about eight days, winter in the northern hemisphere being eight days shorter than summer. In the case of Mars, owing to the much greater eccentricity of his orbit combined with his longer period, the difference amounts to 74 days. In one hemisphere winter is long and cold, summer short and hot; in the other winter and summer interchange. Owing to the present position of the line of apsides, the line connecting the points of Mars' nearest approach to and farthest recession from the Sun, the former hemisphere happens to be the southern one; the latter, the northern. The lengths of their respective seasons are as follows:--

In the northern hemisphere, winter lasts 147 of his own days; spring 191 days; summer, 181 days; autumn, 149 days; while in the southern hemisphere, winter lasts 181 days; spring, 149 days; summer, 147 days; autumn, 191 days.

Curiously enough, an analogous distribution of heat and cold occurs also at the present time in the case of the Earth; its axis and line of apsides holding the same relation to each other that the Martian ones do. This similarity of aspect is, as we shall see later, apparently very curiously reproduced in certain peculiarities of the surfaces of the two planets. But with Mars the result is much more marked on account of the greater eccentricity of his orbit, which is .0931 as against the Earth's .0168.

As even under these exaggerated conditions his two polar regions show much alike, modern theories about our glacial epochs are considerably shaken. The last of the preliminary points to be taken up is the form of the planet. Consideration of it makes in some sort a bridge from the planet's past to its present. For its deviation from a perfect sphere tells us something of its history.

Between the shapes of the large planets, Jupiter, Saturn, Uranus, and probably Neptune, and those of the small ones, Mercury, Venus, the Earth, and Mars, there is a striking dissimilarity, the former being markedly oblate spheroids, the latter almost perfect spheres.

Into the cause of this, very interesting as it is, we have not here space to go. The effect, however, is so noticeable that while the most casual glance at the disk of Jupiter will reveal its ellipticity, the most careful scrutiny would fail to show Mars other than perfectly round.

Nevertheless, the planet is slightly flattened at the poles. Measures have repeatedly been made to determine the extent of this flattening, with surprisingly discordant results, most of the values being much too large.

Observations at Flagstaff during this last opposition have not only shown that most of the values were too large, but have revealed the cause of their discrepancy. There turns out to be a factor in the case, hitherto unsuspected, whose presence proves to be precisely such as would cause the observed variations in measurements. It not only accounts for the fact of discrepancy, but for the further fact that the discrepancies should usually be on the side of an increase of the apparent polar flattening. This factor is the recognition of a perceptible twilight upon the planet, not only of enough account to be visible, but to have been actually measured, quite unconsciously, by Mr. Douglass, and disclosed only when the measures came to be compared with each other. Of this I shall speak more at length when we reach the subject of atmosphere. Here it is only necessary to say that the presence of a twilight fringing the surface of the planet would have the effect of increasing the apparent size of the equatorial diameter at all times, but to a different degree at different times, and almost always more than it would the polar one. In consequence, the