liv PROCEEDINGS OF THE GEOLOGICAL SOCIETY. 



essentially geometrical, and so far are independent of the physical 

 causes to which the phsenomena may be referred. The proofs of 

 them must necessarily rest on observation independently of any in- 

 vestigation of the operation of mechanical agencies, and might be 

 complete without any physical hypothesis whatever if the theories 

 were sufficiently exact. But where approximate proofs are all which 

 the nature of the case will admit of, the probable truth of the theo- 

 ries maj^ be increased, if it can be shoA^Ti that the laws which, accord- 

 ing to the theories, characterize the phsenomena would be the neces- 

 sary results of some probable mechanical agency. 



Regarding the subject in this point of view, the question naturally 

 suggests itself — whether phsenomena of elevation characterized by 

 any law of symmetry, like that for instance of the pentagonal reseau, 

 would result from any assignable physical cause acting under pro- 

 bable conditions ? 



In answering this question I shall be led to make a few remarks 

 on the physical theory adopted by M. de Beaumont, but vnthout 

 any intention of entering into a full discussion of it. The cause to 

 which this theory assigns the phsenomena of elevation is the contrac- 

 tion of the earth's mass in consequence of its refrigeration. It is 

 assumed that the primitive state of the earth was one of entire 

 fluidity, and that it consists at present of a solidified shell, extremely 

 thin compared with the earth's radius, and an enormous internal 

 fluid nucleus. The temperature of the incipient solidified shell would 

 decrease faster than that of the interior nucleus, and would therefore 

 tend to contract more rapidly than the nucleus. During the earliest 

 stages therefore of solidification, the shell would be in a state of ten- 

 sion unless relieved by the formation of cracks and fissures. After a 

 certain time, however, which M. de Beaumont considers to have 

 long since expired, the converse would hold. The fluid nucleus would 

 contract more rapidly than the solid crust, and would consequently 

 tend to separate from it, and thus leave it unsupported beneath. In 

 such case, if the crust were sufficiently thin, it might be unable to 

 support itself, and might therefore collapse by its weight, and we may 

 conceive ridges and furrows to be thus produced on its surface, con- 

 stituting longitudinal lines of elevation. M. de Beaumont supposes 

 these collapses to take place suddenly at distinct epochs, and a 

 distinct system of synchronous parallel lines of elevation to be formed 

 at each collapse. 



There can be no doubt, I conceive, that the contraction of the 

 earth's solid crust arising from the exceedingly slow diminution of 

 the supporting fluid pressure beneath, would, in a great measure, be 

 also slow and continuous. Whether this would be attended with 

 sudden and violent collapses at long intervals of time is extremely 

 uncertain. I believe this hypothesis to be more favourable to the 

 theory which would assign phsenomena of elevation to slow and con- 

 tinuous, or to small and often repeated movements, than to that 

 which attributes them to movements of a paroxysmal character. I 

 have no intention, however, of discussing this general question, for 

 which we have at present but very imperfect data. Admitting, then, 



