THE MOON 393 



Now, until the water in the cistern has evaporated 

 down to the level of the first tread (beginning from the 

 top), the falling snow cannot gain a footing on the water 

 surface, but must settle only on the brim and a slight dis- 

 tance beyond, its depth gradually decreasing to our pre- 

 scribed snowfall limit. A yard's depth of water, area for 

 area, being equivalent to ten times that heighth of snow, 

 we shall have gained by the evaporation of the first yard 

 a snow rampart all around the brim, say five feet high, 

 sloping gradually backward to nothing. I have just 

 stated that the snow could settle only on the brim, but this 

 is not precisely true, for we must not forget the tallest 

 of the granite blocks, which for one-third of this time has 

 been exposed. It, too, is capped with snow, say to the 

 depth of two feet. 



Now suppose another equal period to elapse, and the 

 cistern's depth reduced to the level of the second tread. 

 This time the brim's cap has increased by another five 

 feet in heighth, and the slope behind it in slightly less pro- 

 portion, because the rising snow brim has in a measure 

 further blocked the spread of the snow. Again, the first 

 tread having all this time been clear of the water, it, too, 

 is covered, and to a depth of five feet ; the tallest granite 

 island has also gained six more feet in height; and the 

 second one of eight feet, having been above water two- 

 thirds of the period, disports a snow cap of four feet. 

 Thus the process goes on till we reach the circular area at 

 the bottom. In this area we have finally four islands 

 standing in a shallow pool of water less than a yard deep. 

 But note that the first block uncovered is not the tallest 

 now, for its pinnacle being, let us say, only two feet 

 square, lacks the area necessary to sustain a taller col- 

 umn. As a matter of fact the third block has much the 

 broadest top of any, and though still shorter than either 

 the first or second, must in time, as the result of future 

 processes of evaporation and snowfall, grow to a great 

 height; yet, being itself limited in area also, it can per- 

 haps never reach the altitude of the encircling rampart, 

 which by now, let us say, has risen to the imposing alti- 



