After it had been found that, to ordinary tests, the moon offered no evidence of the possession of an atmosphere, and before Dr. Stoney's theory was broached, it was supposed by many that the moon had lost its original supply of air by absorption into its interior. The oxygen was supposed to have entered into combination with the cooling rocks and minerals, thus being withdrawn from the atmosphere, and the nitrogen was imagined to have disappeared also within the lunar crust. For it seems to have always been tacitly assumed that the phenomenon to be accounted for was not so much the _absence_ of a lunar atmosphere as its _disappearance_. But disappearance, of course, implies previous existence. In like manner it has always been a commonly accepted view that the moon probably once had enough water to form lakes and seas. These, it has been calculated, could have been absorbed into the lunar globe as it cooled off. But Johnstone Stoney's theory offers another method by which they could have escaped, through evaporation and the gradual flight of the molecules into open space. Possibly both methods have been in operation, a portion of the constituents of the former atmosphere and oceans having entered into chemical combinations in the lunar crust, and the remainder having vanished in consequence of the lack of sufficient gravitative force to retain them. But why, it may be asked, should it be assumed that the moon ever had things which it does not now possess? Perhaps no entirely satisfactory reply can be made. Some observers have believed that they detected unmistakable indications of alluvial deposits on lunar plains, and of the existence of beaches on the shores of the "seas." Messrs. Loewy and Puiseux, of the Paris Observatory, whose photographs of the moon are perhaps the finest yet made, say on this subject: "There exists, from the point of view of relief, a general similarity between the 'seas' of the moon and the plateaux which are covered to-day by terrestrial oceans. In these convex surfaces are more frequent than concave basins, thrown back usually toward the verge of the depressed space. In the same way the 'seas' of the moon present, generally at the edges, rather pronounced depressions. In one case, as in the other, we observe normal deformations of a shrinking globe shielded from the erosive action of rain, which tends, on the contrary, in all the abundantly watered parts of the earth to make the concave surfaces predominate. The explanation of this structure, such as is admitted at present by geologists, seems to us equally valid for the moon."[17] [Footnote 17: Comptes Rendus, June 26, July 3, 1899.] It might be urged that there is evidence of former volcanic activity on the moon of such a nature that explosions of steam must have played a part in the phenomena, and if there was steam, of course there was water.