92 THE GREAT LUNAR CRATER TYCHO. 



that the moon was yradaally formed, tlie suuill bodies constitntiiii'' the 

 ring liaviiiii' iirst coalesced into a large nnniber of inoonlets wliich tinally 

 all united into a single sphere. According to this hypothesis the Innar 

 craters are the scars produced by the collision of the nioonlets which 

 last surrendered their individuality, and according to Mr. Gilbert and 

 a mathematical friend who aided him in the investigation, 58 per cent 

 of the moonlets would, under the circumstances imagined, strike the 

 surface of the moon, making an angle of less than 20° with the vertical, 

 while 70 per cent would strike at an angle of less than 30°, and 80 per 

 cent at an angle of less thau 40°. From laboratory experiments Mr. 

 Gilbert tbund that the ellipticity of the scars ou his jilastic target 

 increased slowly up to an incidence of 40° to the vertical, and that 

 beyond that incidence the resulting scars showed considerable ellip- 

 ticity. He assumes that, owing to the flat character of the Saturniau 

 ring abont the earth, the moonlets must have approached the moon 

 approximately in the plane of its e(piator, but the fact is not attested 

 by the grouping of the craters in a medial zone. Mr. Gilbert therefore 

 assumes that the axis of the moon's rotation has shifted under the suc- 

 cessive impulses of the bombardment, and that the moon's equator has 

 occupied successively all i)arts of its surface. He assumes that the 

 velocity of impact due to the moon's gravity would be sufficient to melt 

 the rocks of the lunar surface, and that they would during a short period 

 behave as if they w^ere composed of plastic material, but would become 

 hardened.before the crater could subside. 



The theory does not at all commend itself to my mind. M. Roche, of 

 Montpelier, showed that a ring abont a planet would break up if it 

 extended beyond a distance of 2 ll-25ths the radins from the center of 

 the planet, and if the density of the planet increased towards its center 

 the maximum limit to which a ring could extend would be still further 

 contracted. A moon formed just outside such a ring would have au 

 ellipticity greater than that of an ordinary hen's egg; and as tidal 

 action carried the moon away from its primary it would gradually 

 approximate to a s])herical form. One can hardly conceive that such a 

 change of shape could take place without obliterating scars on its sur- 

 face; but there is another objection to the theory, which, to my mind, 

 is even more conclusive. There are upon the moon many lines or 

 strings of small craterlets which fall very evidently into line with one 

 another. If we are forced to treat them as scars upon a target, we 

 must regard their allineation as the result of meie chance distribution; 

 but the number of such strings i)recludes any such assumption; there 

 must therefore be a physical reason for the allineation, and the nu)st 

 obvious assumption seems to be that the craterlets mark out a line of 

 weakness in the crust of the moon and lie along a volcanic fissure or 

 Innar fault. 



There is ever gradation in size and in tyi)e from the small craterlets 

 or cup-shaped depressions up to the gigantic walled rings, and any 



