An examination of such enlarged photographs, which permits us to embrace with the eye a large surface, filled with a mass of naturetrue details, has led MM. Loewy and Puiseux to some interesting suggestions concerning the origin of the so-called 'rills' or groups of parallel rents in the Moon's crust.5 And on the other side of the Atlantic, the direct observations lately made by Professor W. Pickering under the clear sky of Peru, as well as his studies of the American photographs, have produced such new data concerning the atmosphere of the Moon, and the possible existence of water on its surface, as are sure to give a quite fresh interest to lunar studies."

The Moon is so small in comparison with the Earth (its weight is eighty-one times less), and consequently the force of gravity is so much smaller on its surface, that, even if it had an atmosphere of the same composition as ours, its density in its lowest parts would be from thirty to fifty times less than the density of our atmosphere at the sea-level. But it appears from Dr. Johnstone Stoney's 7 investigations, that even if the Moon was surrounded at some time of its existence with a gaseous envelope consisting of oxygen, nitrogen, and water vapour, it would not have retained much of it. The gases, as is known, consist of molecules, rushing in all directions at immense speeds; and the moment that the speed of a molecule which moves near the outward boundary of the atmosphere exceeds a certain limit (which would be about 10,600 feet in a second for the Moon), it can escape from the sphere of attraction of the planet. Molecule by molecule the gas must wander off into the inter-planetary space; and, the smaller the mass of the molecule of a given gas, the feebler

the same detail on two or three negatives settles all possible doubts as to its reality. (L. Weineck and E. S. Holden, 'Selenographical Studies' in Publications of the Lick Observatory, 1894, vol. iii.; Loewy and Puiseux, in Comptes Rendus, tome cxix. p. 254, and tome cxxi. pp. 6, 79; Folie, Bulletin of the Belgian Academy, 1895, vol. xxix. No. 1; and Dr. Klein in Sirius, 1895, p. 112.)

4 Comptes Rendus, 8 juillet 1895, tome cxxi. p. 79.

To explain the origin of these rents, Loewy and Puiseux look for the time when the rocks were in an igneous half-liquid state and floating islands of consolidated scoria were formed on the surface of the molten rocks and drifted like the ice-floes in the Arctic Ocean. Remaining in that sphere of ideas, it may, however, be remarked that the same rents might have originated when the whole crust was already solidified. When Lake Baikal is covered with a thick sheet of ice, and the level of the water goes slightly down in the winter, the ice is intersected by long rents, one to ten yards wide, which usually appear in about the same places and in the same directions. They run in straight lines, have vertical sides, and when the water at their bottom is frozen, they become miniature models of lunar rents.

• William H. Pickering, 'Investigations in Astronomical Photography,' in Annals of the Astronomical Observatory of Harvard College, vol. xxxii. part 1 (Cambridge, Mass., 1895). See also Dr. Klein's analysis of the same (Sirius, 1895, Hefte 7, 8, und 9).

·

On the Physical Constitution of the Sun and Stars,' in Proceedings of the Royal Society for 1868; and paper On the Cause of Absence of Hydrogen from the Earth's atmosphere, and of Air and Water from the Moon,' read on the 20th of April, 1892, before the Royal Dublin Society.

the planet's attraction, and the higher the temperature at the boundary of its atmosphere, the sooner the escape of the gas must take place. This is why no free hydrogen could be retained in the Earth's atmosphere, and why the Moon could retain no air or water vapour.

However, neither these speculations, which are very likely to be true, nor Bessel's previous calculations, could convince practical astronomers of the absolute absence of any atmosphere round the Moon. A feeble twilight is seen on our satellite, and twilight is due, as is known, to the reflection of light within the gaseous envelope; besides, it had been remarked long since at Greenwich that the stars which are covered by the Moon during its movements in its orbit remain visible for a couple of seconds longer than they ought to be visible if their rays were not slightly broken as they pass near to the Moon's surface. Consequently, it was concluded that the Moon must have some atmosphere, perhaps only 200 times thinner than our own. Of course, a gaseous envelope so thin as that would only be noticeable in the deeper valleys, and it would attain its greatest density within the circus-like cavities whose floor, as a rule, lies deeper than the surrounding country. Towards the tops of the mountains it would be imperceptible. But nevertheless, as was shown by Neison, it would play an important part in the economy of life on the Moon's surface.

The observations made at Lick, at Paris, and at Arequipa, fully confirm this view. A twilight is decidedly visible at the cusps of the crescent-moon, especially near the first and the last quarter. It prolongs the cusps as a faint glow over the dark shadowed part, for a distance of about seventy miles (60"), and this indicates the existence of an atmosphere having on the surface of the Moon the same density as our atmosphere has at a height of about forty miles. A similar result is obtained when the slight flattening of the disc of Jupiter, which takes place when the planet is just going to be covered by the Moon, or emerges from behind it, is measured on the Arequipa photographs. Such an atmosphere is next to nothing, but there is another observation, namely, of a dark band appearing between Jupiter and the Moon's limb when the former begins to be covered by the latter; and Professor Pickering finds no other explanation for it than in some very light haze, partly due to water vapour, which would rise a few miles above the Moon's surface where it is illuminated by the rays of the Sun.

Such a supposition would have been met some time ago with great suspicion. But it must be said that the more the Moon's surface is studied in detail the more astronomers are inclined to think that, in some places at least, a haze, originated from water vapour, is the only possible means to explain certain curious occurrences. Thus, Dr. Sarling has lately reminded us that, in 1774, Eysenhard, a pupil of Lambert, saw the part of the shadow line which crossed one

of the plains (the Mare Crisium) brought in a wave-like movement which lasted for two hours and was seen by three different persons— only in this part of the lunar disc. Those undulations, which spread at a speed of 1,200 feet per second over a distance of eighty miles, could only be due-as Dr. Sarling truly remarks-to vapours floating over the plain. In several instances, the interiors of deep lunar circuses took a misty appearance at sunrise, and this misty appearance disappeared as the Sun rose higher above the same circus, while in other cases it persisted a considerable time after sunrise, even though all around was sharply marked and distinct. And so on. The temperature of the Moon's surface, when it is heated by the Sun's rays, being very near to the freezing point, as appears from Langley's last measurements, the evaporation of frozen water under the rays of the rising Sun is surely not at all improbable.

It remains, of course, to be seen whether a haze of this sort is not due in some cases to water ejected by volcanoes or geysers; the more so as some volcanic activity, remodelling until now the forms of the craters, seems to exist. There is, indeed, among astronomers a strong suspicion of a lunar crater, nearly three miles in diameter, being of recent formation. It was first discovered by Dr. Klein in 1876, in the plain named Mare Vaporum, after he himself and many others had previously so often examined that region without seeing the crater. Besides, the alternate appearance and disappearance of another crater (Linné), nearly four miles in diameter, can hardly be explained unless it is concealed from time to time by the vapours which it itself exhales. As to changes observed in the shapes of small lunar volcanoes, they are too numerous to be due to mere errors of observation, 10

If free water thus exists occasionally, even now, on the Moon's surface, or has existed at a relatively recent period, it is natural to ask whether it has left no traces of its activity. Are there no rivervalleys which would bear testimony to its existence? Till lately, the majority of astronomers answered this question in the negative, even though their earlier predecessors, armed with feebler telescopes, were most affirmative on this point. The maria, or seas, are known to be plains on which no traces of aqueous action have been detected, and the clefts, or large 'rills,' are almost certainly rents produced in a solid surface.

However, beside these clefts, there are much finer formations which only lately have received due attention, and these finer rills have all the aspects of river-beds. They are not straight-lined, but wind exactly as rivers wind on our maps; they fork like rivers; they

Sirius, 1895, vi. p. 134.

• Edw. Neison, The Moon and the Condition and Configurations of its Surface, p. 33 (London, 1876).

10 The Observatory, June 1892; Nature, vol. xlvi. p. 134.

are wider at one end than at the other, and one end is nearly always higher than the other. Many such fine rills have been observed and mapped lately, and Professor W. Pickering gives a list of thirty-five presumable river-beds, large, medium-sized, and very fine." However, contrary to most terrestrial rivers, the lunar river-beds-those, at least, which were observed by W. Pickering-have their wider end in their upper course, nearly always in a pear-shaped craterlet. This circumstance offers, however, nothing extraordinary, as we know many rivers in Central Asia and South America which originate in a lake and grow thinner and thinner as they enter the arid plains. To take one illusstration out of several, one such river, sixty-five miles in length with all its windings, rises in a craterlet, perhaps 2,000 feet wide, but soon its valley narrows to 1,000 feet, or less, and is lost in a plain. Occasionally such 'rivers' occur in groups on the slopes of the mountains. Other river-beds, on the contrary, seem to have the normal characters of our rivers. One of them begins in the mountains as an extremely fine line, gradually increases in width, and, after having received a tributary, becomes a broad but shallow valley. Another bifurcates into two very fine lines in its higher part.12 In short, it may now be taken as certain that there are river-beds, to all appearance of aqueous origin; but they are so narrow that we should not be able to discover water-courses if they existed at the bottom of these valleys. We must be content with saying that they have been scooped out by running water.

So much having been won, the next step was naturally to ask if no traces of vegetation can be detected. On Mars, we see how every year a snow cover spreads over the circumpolar region, how later on in the season wide channels appear in it, and how the snow thaws gradually-presumably giving origin to water; even clouds have lately been seen; and we can notice, moreover, how the coloration of wide surfaces changes, probably because they are covered with vegetation, and how that coloration gradually takes a reddish yellow tint. Of course, if anything of the sort took place on our nearest neighbour, the Moon, it would have been noticed long since. But it would be most unwise to maintain that nothing similar to it happens, on a much smaller scale. On the contrary, Professor Pickering shows that there are some probabilities in favour of plants of some sort or another periodically growing on the Moon as well.

The great lunar circuses or craters attain, as is known, colossal dimensions; the largest of them have 100 and 130 miles in diameter, and the floor of their inner parts is mostly flat. Now,

"Annals of the Astronomical Observatory of Harvard College, vol. xxxii. part 1, p. 87.

1 Dr. Sarling's letter to Sirius, March 30, 1895; map of the region near Herschel, f, made by J. N. Krieger at the Observatory of Triest, in same periodical, September 1895, p. 195.

Neison had already made the remark that grey, almost black, spots appear on the floor of certain craters at full moon, but disappear later on, and W. Pickering has carefully investigated several such spots during his unfortunately too short stay at Arequipa. Contrary to all expectations, they grow darker just after full moon, that is, when the Sun strikes the visible part of the Moon's surface in full and when it is geometrically impossible for any shadow to be visible, and they become invisible when the Sun is lowest and the shadows are evidently strongest. We know, however, of no stone which would darken under the action of sunlight, and grow lighter when the sunlight fades, and, following two such authorities as Mädler and Neison, Professor Pickering inclines to see the causes of those changes in vegetation. Such spots, whose darkness varies with the Sun's altitude, are not mere accidents. On the contrary, they have been found on all plains, with the exception of one, and in two plains, the Mare Tranquillitatis and Mare Nectaria, they apparently cover the whole floor, their changes being sometimes so conspicuous as to be almost visible to the naked eye. In the craters they always appear in the lower inner edges, but never on the tops of the walls, and rarely, if ever, on the outer walls. As a rule, they are coloured in dark grey, but in one case at least, one of the spots, examined with a great power, was of a 'pronounced yellow colour, with perhaps a suspicion of green.'

These observations,13 which Professor Pickering unhappily found impossible to continue under the much less propitious sky of Massachusetts, on account of the poor quality of the seeing,' are certainly very promising, the more so as they are not isolated. For the last few years, a number of data are accumulating, all tending to prove that it was too rash to describe the Moon's surface as utterly devoid of life. It appears very probable, on the contrary, that volcanic changes continue to go on on the Moon's surface on a larger scale than on the Earth, and that notwithstanding the most unfavourable conditions for organic life which prevail there, such life exists, be it only on a small scale. This is certainly very far from the sanguine affirmations of the last century selenographers, who wanted to see on the Moon fortifications,' 'national roads,' and 'traces of industrial activity;' such objects, if they did exist, could not be seen with our best instruments. But traces of vegetation which develops at certain periods and fades next, traces of water which runs perhaps even now, as well as indications of volcanic changes of the surface, become more and more probable in proportion as we learn to know our satellite better.

18 In the above-mentioned volume of Harvard Annals they are published in full, and are illustrated by a number of excellent photographs.