PT. i. Astronomy and Geology compared. 11 



trace the connection ourselves, and only that, so far 

 as our experience extends, it uniformly follows. 



Mr. Babbage, in his ' Passages from the Life of a 

 Philosopher,' gives an instance of the possible error 

 which may lurk under these arguments, founded on 

 the principle that similar causes will always produce 

 similar effects. He says that he observed that he 

 could set his calculating machine to such a point 

 that a certain combination would be produced a 

 million of times, but at the million-and-first the 

 result would be different. Anyone, he said, merely 

 reasoning from cause and effect would conclude at 

 the 999,000th time that at the million-and-first time 

 the same result would follow which had already 

 preceded so many similar trials ; yet he would be 

 wrong, and his reasoning would be proved to be fal- 

 lacious. Still, although the certainty is not so 

 absolute or the proof so overwhelming as in cases 

 where the reasoning is founded on mathematical 

 demonstration, a degree of proof almost amounting 

 to certainty may be attained through the evidence of 

 facts. We must however admit the possibility of 

 error, and adopt conclusions based upon the relations 

 of cause and effect, with a slight reservation. 



The third method of inquiry does not differ alto- 



