Assured by physical properties that our visual appearances are quite capable of being what they seem we pass to the phenomena of the cap itself. Like as are the polar caps of the two planets at first regard, upon further study very notable differences soon disclose themselves between the earthly and the Martian ones; and these serve to give us our initial hint of a different state of things over there from that with which we are conversant on Earth.

To begin with, the limits between which they fluctuate are out of all proportion greater on Mars. It is not so much in their maxima that the ice-sheets of the two planets vary. Our own polar caps are much larger than we think; indeed, we live in them a good fraction of the time. Our winter snows are in truth nothing but part and parcel of the polar cap at that season. Now, in the northern hemisphere snow covers the ground at sea-level more or less continuously down to 50° of latitude. It stretches thus far even on the western flanks of the continents, while in the middle of them and on their eastern sides it extends ten degrees farther yet during the depth of winter. So that we have a polar cap which is then ninety degrees across. In our southern hemisphere it is much the same six months later, in the corresponding winter of its year.

On Mars at their winter maxima the polar caps extend over a similar stretch of latitude. They do so, however, unequally. The southern one is considerably the larger. In 1903, 136 days after the winter solstice, in the Martian calendar February 27, it came down in longitude 225° to 44° of latitude and may be taken to have then measured ninety-three degrees across; in 1905, 121 days after the same solstice, it stretched in longitude 235° to latitude 42°, and 158 days later, in longitude 221° to latitude 41°; values which, supposing it to have been round, imply for it a diameter on these occasions of ninety-six and ninety-seven degrees. It was then February 20 and March 10 respectively of the Martian year. These determinations of its size at the two oppositions agree sufficiently well considering the great tilt away from us of the south pole at the time and the horizonward foreshortening of the edge of the snow. It seems from a consensus of the measures to have been some five degrees wider in 1903 than in 1905, which may mean a colder winter preceding the former date. The cap was still apparently without a dark contour in both years, showing that it had not yet begun to melt.

Less has been learnt of the northern cap. In 1896-1897 when it was similarly presented skirting the other rim of the disk, a gap occurred in the observations corresponding to the time by Martian months between February 24 and March 22. On the former date the cap came down only to latitude 55° in longitude 352°; on the subsequent one and for several days after the latitude of the southern limit of the snow was such as to imply a breadth to it of about eighty degrees. The cap was now bordered by a dark line, proving that melting had already set in. It cannot, however, at its maximum have covered much more country than this, in view of its lesser extent on February 24.

Fair as our knowledge now is of the dimensions of the Martian polar caps at their maxima, we have much more accurate information with regard to their minima, and this, too, was obtained much earlier. That we should first have known their smallest rather than their greatest extent with accuracy may appear surprising, exactly the opposite being our knowledge of our own. It is not, however, so surprising as it appears, inasmuch as it is an inevitable consequence of the planet’s aspect with regard to the sun. When the tilt of the axis inclines one hemisphere toward the sun, that hemisphere’s polar cap must melt and dwindle, while at the same time it is the one best seen, the other being turned away from the sun and therefore largely from us as well; so that even such part of the latter as is illumined lies low down toward the horizon of the disk where a slight change of angle means a great difference in size.

It has thus come about that both the south and the north polar caps have been repeatedly well seen and measured at their minimum; and the measures for different Martian years agree well with one another. For the northern cap six degrees in diameter is about the least value to which it shrinks. The south one becomes even smaller, being usually not more than five degrees across, while in 1894 it actually vanished, a thing unprecedented. Its absence was detected by Douglass at Flagstaff and shortly after the announcement of its disappearance the fact was corroborated by Barnard at the Lick. The position the cap would have occupied was at the time better placed for observation in America than in Europe, inasmuch as the cap is eccentrically situated with regard to the geographic pole and its centre was then well on the side of the disk presented to us while in Europe it was turned away. This, together with the fact that it undoubtedly came and went more than once about this time, accounts for its disappearance not having been recognized there, haze left by it having apparently been mistaken for the cap itself.

On Earth the minima are much larger. In the northern hemisphere the line of perpetual snow or pack-ice in longitude 50° east runs about on the 80° parallel, including within it the southern end of Franz Joseph Land. Opposite this, in longitude 120° west, above the North American continent, it reaches down lower still to 75°. So that the cap is then from twenty to thirty degrees in diameter. In the southern hemisphere it is even larger. In longitude 170° west the land was found by Ross to be under perpetual snow in latitude 72°. Cook had reached in longitude 107° east an impassable barrier of ice in latitude 70° 23'. The season was then midsummer, January 30. So that we are perhaps justified in considering 71° south as about the average limit of perpetual snow or paleocrystic ice. This would make the southern cap at its minimum thirty-eight degrees across. Pack-ice with open spots extends still farther north. The Pagoda in 1845 was stopped by impenetrable pack-ice in south latitude 68° and the Challenger in 1874 encountered the pack in latitude 65° on the 19th of February, which corresponds about to our 19th of August, the time at which the sea should be most open. The limit of perpetual snow is thus lower in the southern than in the northern hemisphere. Here again, then, the two minima differ, but in the reverse way from what they do on Mars.

From this we perceive that the variations in size of the caps are much more striking on Mars than on the Earth and that these are due chiefly to the difference in the minima, the maxima not varying greatly.

To explain these interesting diversities of behavior in the several polar caps we shall have to go back a little in general physics in order to get a proper take off. It is a curious concomitant of the law of gravity that the amount of heat received by a planet in passing from any point of its path to a point diametrically opposite is always the same no matter what be the eccentricity of the orbit. Thus, a planet has as many calories falling upon it in travelling from its vernal equinox to its autumnal as from the autumnal to the vernal again, although the time taken in the one journey be very different from that of the other. This is due to the fact that the angle swept over by the radius vector, that is, the imaginary bond between it and the sun, is at all points proportional to the amount of heat received; just as it is of the gravity undergone, the two forces radiating into space as the inverse square of the distance. Thus the heat received by a point or a hemisphere, through any orbital angle, is independent of the eccentricity of the orbit.

But it is not independent of the axial tilt. For the force of the sun’s rays is modified by their obliquity. The amount of heat received at any point in consequence of the tilt turns upon the position of the point, and for any hemisphere taken as a whole it depends upon the degree to which the pole is tilted to the source of heat. In consequence of being more squarely presented to its beams, the hemisphere which is directed toward the sun and therefore is passing through its summer season gets far more insolation than that which is at the same time in the depth of its winter. For a tilt of twenty-four degrees, the present received value for the axis of Mars, the two hemispheres so circumstanced get amounts of heat respectively in the proportion of sixty-three to thirty-seven.

But, though the summer and winter insolation thus differ, they are the same for each hemisphere in turn. Consequently the mere amount of heat received cannot be the cause of any differences detected between the respective maxima and minima of the two polar caps. If heat were a substance which could be stored up instead of being a mode of motion, the effect produced would be in accordance with the quantity applied and the two caps would behave alike. As it is the total amount has very little to say in the matter.

Not the amount of heat but the manner in which this heat is made at home is responsible for the difference we observe. Now, though the total amount is the same in passing from the vernal to the autumnal equinox as from the autumnal to the vernal, the time during which it is received in either case varies from one hemisphere to the other. It is summer in the former while it is winter in the latter and the difference in the length of the two seasons due to the eccentricity of the orbit makes a vast difference in the result. Winter affects the maxima, summer the minima, attained. Of these opposite variations presented to us by the two caps, the maxima, the one most difficult to detect, is the easiest to explain, for the difference in the maxima seems to be due to the surpassing length of the antarctic night.

Owing to the eccentricity of the orbital ellipse pursued by Mars and to the present position of the planet’s solstices, the southern hemisphere is farther away from the sun during its winter and is so for a longer time. The seasons are in length, for the northern hemisphere: spring, 199 days; summer, 183 days; autumn, 147 days; and winter, 158 days; while for the southern hemisphere they are: spring, 147 days; summer, 158 days; autumn, 199 days; and winter, 183 days. The arctic polar night is thus 305 of our days long; the antarctic, 382. Thus for 77 more days than happens to its fellow the southern pole never sees the sun. Now, since the total sunlight from equinox to equinox is the same in both hemispheres, its distribution by days must be different. In the southern hemisphere the same amount is crowded into a smaller compass in the proportion of 305 to 382; that being that hemisphere’s relative ratio of days. But since during winter the cap increases, there is a daily excess of accumulation over dissipation of snow and each twenty-four hours must on the average add its tithe to the sum total. Since the northern days are the warmer each adds less than do the southern ones; and furthermore there are fewer of them. On both these scores the amount of the deposition about the northern pole should be less than about the southern one. Consequently, the snow-sheet there should be the less extensive and show a relatively smaller maximum, which explains what we see.

With the minima the action is otherwise. Inasmuch as the greater heat received during the daylight hours by the southern hemisphere is exactly offset by the shortness of its season, it would seem at first as if there could be no difference in the total effect upon the two ice-caps.

But further consideration discloses a couple of factors which might, and possibly do, come in to qualify the action and account for the observed effect. One is that though the total amount of heat received is the same, the manner of its distribution differs in the two hemispheres. In the northern one the time from vernal to autumnal equinox is 382 days against 305 in the southern. Consequently, the average daily heat is then five fourths more intense in the southern hemisphere. Indeed, it is even greater than this and nearer four thirds, because the melting occurs chiefly in the spring and in the first two months of summer when the contrast in length of season between the two hemispheres is at its greatest. Now, a few hotter days might well work more result than many colder ones. And this would be particularly true of Mars where the mean temperature is probably none too much above the freezing-point to start with. Ice consumes so much caloric in the process of turning into any other state, laying it by in the form of latent heat before it can turn into water and then so much more before this water can be converted into steam that a good deal has to be expended on it before getting any perceptible result. Once obtained, however, the heat is retained with like tenacity. So that the process works to double effect! If sufficient heat be received the ice is first melted, then evaporated and finally formed into a layer of humid air, the humidity of which keeps it warm. Dry air is unretentive of heat, moist air the opposite. And for the melting of the ice-cap to proceed most effectively the temperature that laps it about must be as high as possible and kept so as continuously as may be. If between days it be allowed to fall too low at night much caloric must needs be wasted in simply raising the ice again to the melting-point. This a blanket of warm air tends to prevent, and this again is brought about by a few hot days rather than by many colder ones. It is not all the heat received that becomes effective but the surplus heat above a certain point. The gain in continuity of action thus brought about is somewhat like that exhibited between the running of an express and an accommodation train. To reach its destination in a given time the former requires far less power because it does not have to get up speed again after each arrest. Thus the whole effect in melting the snow would be greater upon that hemisphere whose summer happens to be the more intense.

The greater swing in size of the cap most exposed to the effects of the eccentricity is, then, the necessary result of circumstances when the precipitation is not too great to be nearly carried off by the subsequent dissipation. This is the state of things on Mars and the second of the factors above referred to. On the Earth as we have seen the polar caps are somewhat larger at their maximum and very much so at their minimum. Now, this is just what should happen were the precipitation increased. Suppose, for example, that the amount of precipitation were to increase while the amount of summer melting remained the same, and this would be the case if the vapor in the air augmented for one cause or another, and the result of each fresh deposit was locked up in snow. After a certain point the cap would grow in depth rather than in extension; the winter deposit would be thicker but the summer evaporation would remain the same. Now, if this occurred, it is evident that the minimum size of the cap would increase relatively much faster than the maximum, and furthermore, that the relative increase of the minimum in the two caps would be greatest for that which had seasons of extremes. The result we see in the case of the Earth. In the arctic cap, where in consequence of the eccentricity of the orbit the winter is shorter, the maximum is less than in the antarctic and this extra amount of precipitation cannot be wholly done away with in its intenser summer, so that the minimum too is greater there.

We reach, then, this interesting conclusion. We find that eccentricity of orbit by itself not only causes no universal glaciation in the hemisphere which we should incidentally suppose likely to show it, but actually produces the opposite result, in more than offsetting by summer proximity what winter distance brings about. To cause extensive glaciation we must have, in addition to favorable eccentricity, a large precipitation. With these two factors combined we get an ice age, but not otherwise. The result has an important bearing on geologic glacial periods and their explanation.

Once formed, an ice-sheet cools everything about it and chills the climate of its hemisphere. It is a perpetual storehouse of cold. Mars has no such general glaciation in either hemisphere, and the absence of it, which is due to lesser precipitation, together with the clearness of its skies, accounts for the warmth which the surface exhibits and which has been found so hard hitherto to interpret. Could our earth but get rid of its oceans, we too might have temperate regions stretching to the poles.