To what, then, was the limb-light due? At first sight, it would seem as if the Moon might help us; for the Moon's rim is similarly ringed by a lune of light. In her case the effect has been attributed to mountain slopes holding the Sun's light at angles beyond the possibility of plains. But Mars has few mountains worthy the name. His terminator--that is, the part of the disk which is just passing in or out of sunlight, and discloses mountains by the way in which they catch the coming light before the plains at their feet are illuminated--shows irregularities quite inferior to the lunar ones, proving that his elevations and depressions are relatively insignificant.

Not due, then, to either mountains or mist, there is something we know that would produce the effect we see,--dust or water particles in the Martian air; that is, just as the Earth's atmosphere is somewhat of a veil, so is the Martian one, and this veiling effect, though practically imperceptible in the centre of the disk, becomes noticeable as we pass from the centre to the edge, owing to the greater thickness of the stratum through which we look. At thirty degrees from the edge, our line of sight pierces twice as much of it as when we look plumb down upon the centre of the disk and more yet as we approach the edge itself; in consequence, what would be diaphanous at the centre might well seem opaque toward the limb. The effect we are familiar with on Earth in the haze that always borders the horizon,--a haze most noticeable in places where there is dust, or ice, or water in the air. Here, then, we have a hint of the state of things on Mars. Ice particles both are probable and would give the brilliancy required.

This first hint receives independent support from another Martian phenomenon. Contrary to what the distance of the planet from the Sun and the thinness of its atmospheric envelope would lead us to expect, the climate of Mars appears to be astonishingly mild. Whereas calculation from distance and atmospheric density would put its average temperature below freezing, thus relegating it to perpetual ice, the planet's surface features imply that the temperature is relatively high. Observation gives every evidence that the mean temperature must actually be above that of the Earth; for not only is there practically no sign of snow or ice outside the frigid zone at any time, but the polar snow-caps melt to a minimum quite beyond that of our own, affording rare chance for quixotic polar expeditions. Such pleasing amelioration of the climate must be accounted for, and aqueous vapor seems the most likely thing to do it; for aqueous vapor is quite specific as a planetary comforter, being the very best of blankets. It acts, indeed, like the glass of a conservatory, letting the light-rays in and opposing the passage of the heat-rays out.

The state of things thus disclosed by observation, the cloudlessness and the rim of limb-light, turns out to agree in a most happy manner with what probability would lead us to expect; for the most natural supposition to make a priori about the Martian atmosphere is the following: When each planet was produced by fission from the parent nebula, we may suppose that it took with it as its birthright its proportion of chemical constituents; that is, that its amount of oxygen, nitrogen and so forth was proportional to its mass. Doubtless its place in the primal nebula would to a certain extent modify the ratio, just as the size of the planet would to a certain extent modify the relative amount of these elements that would thereupon enter into combination. Supposing, however, that the ratio of the free gases to the other elements remained substantially the same, we should have in the case of any two planets the same relative quantity of atmosphere. But the size of the planet would entirely alter the distribution of this air.

Three causes would all combine to rob the smaller planet of efficient covering, on the general principle that he that hath little shall have less.

In the first place, the smaller the planet the greater would be its volume in proportion to its mass, because the materials of which it was composed, being subjected to less pressure owing to a lesser pull, would not be crowded so closely together. This is one reason why Mars should have a thinner atmosphere than our Earth.

Secondly, of two similar bodies, spheres or others, the smaller has the greater surface for its volume, since the one quantity is of two dimensions only, the other of three. An onion will give us a good instance of this. By stripping off layer after layer we reach eventually a last layer which is all surface, inclosing nothing. We may, if we please, observe something analogous in men, among whom the most superficial