358 FROM NEBULA TO NEBULA 



The rills and "canals" are dried-up streams, con- 

 necting straits, or, in some cases, arms of the seas, that, 

 should the snow all melt away entirely, would fill with 

 water to the brim. The reason why they never fill up 

 with snow is surely not hard to understand. Though the 

 sun's heat may fail to bare the ground of snow univer- 

 sally, it is not said that he does not thaw great quantities 

 of it away, or, even, that some of the land may not be 

 periodically bared in exceptional places. The water from 

 these thaws, like that from rain, seeks its lowest level and, 

 as a matter of course, drains into the ancient channels 

 and pools. Arriving there it dissolves what snow may 

 have previously settled in them, or tumbled into them 

 from the icy cliffs on either side, and then, being cor- 

 nered, as it were, by the sun, it rapidly starts evaporat- 

 ing, and the vapor, changing promptly back into snow, 

 settles down again and systematically restores the walls 

 lining the banks. 



Owing to the circumstance that the lunar atmosphere 

 has now found burial in the snow mounds, there are no at- 

 mospheric storms on the satellite, and the freshly-formed 

 snow does not scatter much but settles on the cliffs closest 

 by. Wherever the natural stream or pool is shallow, 

 therefore, the lower are the adjacent walls, and, converse- 

 ly, the deeper these are the more imposing the snow 

 phalanxes hemming them in. On this hypothesis, then, 

 there is no enigma in the phenomenon, otherwise inexpli- 

 cable, that these supposed "fissures in the lunar crust " 

 cleave straight through the very ruggedest of mountains 

 and abhor the valleys ; for this is precisely what should be 

 expected of them. Nor is there any greater marvel in the 

 observed fact that such clefts and the pits of the deeper 

 craters actually delve down deep below the natural level 

 of the surface, thereby adding to the ' ' cavernous ' ' effects. 

 In terrestrial volcanoes the floor of the crater, far from 

 being deeper down than the surrounding regions, is in- 

 variably much higher a most significant distinction. 



