272 
GILBERT. 
they settled downward and their lower portions flowed in¬ 
ward toward the center of the cup. The inward flow from 
all sides produced at the center an upward movement, occa¬ 
sioning the central hill. The effect was perhaps heightened 
by the elastic recoil of a considerable tract of the moon’s 
mass below and about the point of impact. At the same 
time the fused parts, which were partly determined by the 
distribution of strains and partly by the occurrence of local 
passages of more fusible material, flowed to the bottom of 
the cup, either surrounding the central hill or, if in great 
volume, submerging it. Sometimes minor tracts of fused 
matter occurred in the wreath, and the exudation of these 
gave rise to lava streams flowing down the outer slope. The 
inward flow of the lower portions of the walls undermined 
the upper portions, including the inner part of the wreath, 
so that they settled down toward and into the liquid pool 
of the interior, and this settling gave rise to the inner cliff 
and the inner terraces. In the case of some of the large 
craters all of the wreath was carried down. 
The effect of the collision on the moonlet was not uniform 
throughout. The part in immediate contact with the moon, 
being compressed by the shock of the entire mass behind it, 
was probably heated more than any other part. The oppo¬ 
site portion of the moonlet, sustaining no blow from behind 
and having its motion arrested in a comparatively gradual 
way, was less affected and probably never fused; the results 
of laboratory experiment indicate that it remained central in 
the crater and was uplifted by the recoil so as to constitute 
the surface of the central hill. 
The impact theory as thus developed appears competent 
to explain the origin of all typical features of the lunar cra¬ 
ters. Its relation to exceptional features, as well as to asso¬ 
ciated phenomena, will presently be considered; but some¬ 
thing should first be said with reference to certain physical 
factors of the process which are somewhat unfamiliar. 
The production of heat by impact is a well known phe¬ 
nomenon, but instances in which that heat suffices to pro- 
