i9o6.] SEE-THE CAUSE OF EARTHQUAKES. 311 



arise from foldings of the crust, and as the earth's crust is thick, a 

 zig-zag form with many bends in it would be improbable, because a 

 crack started in the rocks anywhere would be likely to run in a 

 straight line. There may be some justice in this criticism, but it 

 seems quite fully compensated for by the fact that the chain actually 

 bends wherever the coast line alters its course. Thus in practice 

 the chain is shown to be capable of flexure wherever the shore line 

 changes its direction, and there does not seem anything improbable 

 in many short bends, unless the trend has some connection with the 

 coast. Whether the chain could fairly be conceived as bending at 

 such short intervals is a question we need not enter into, for we may 

 observe that short bends actually appear in certain chains, and thus 

 the hypothesis is not contrary to nature under certain conditions. In 

 the present case we have made no assumption as to causes, except 

 that the lay of the chain is independent of the seacoast, and hence 

 the hypothesis postulates nothing improbable. 



To reduce the first method to numbers, we may observe that the 

 length of the chain of the Andes is 4,400 miles and the average 

 distance from the sea about 66 miles. Thus 



4,400 I 



n= — ^- = 66^ or P= -. — ^- = 1 : (decillion)^ 

 56 ^ (10)^^ ^ ^ 



If other parallel ranges be included, this divisor would be much 

 increased, perhaps nearly squared. On the other hand, the method 

 of viewing the chain as a whole makes the divisor a quantity of 

 the order of one hundred billions. The truth must, I think, lie 

 somewhere between these extremes. If we were to take account also 

 of the mountains parallel to the Atlantic coast, it would certainly be 

 moderate to conclude that the parallelism to the seashore noticed 

 in the whole of South America, so far as it depends on chance, would 

 be less than i : (decillion)^. 



Thus it is clear that the probability of a physical cause connect- 

 ing the mountains with the parallel shore of the sea is probably 

 more than a decillion decillions to unity, and certainly more than 

 one hundred billions to unity. 



The parallelism noticed in North America along the Pacific and 

 Atlantic coasts, is not less pronounced, nor is there less extent of 



PROG. AMER. PHIL. SOC., XLV. 184 T, PRINTED FEBRUARY 20, I507. 



