OUR CHIEF TIME-PIECE LOSING TIME. 43 



is issued a thick octavo volume crowded with such 

 predictions, three or four years in advance of the 

 events predicted ; and these predictions are accepted 

 with as little doubt by astronomers as if they were 

 the records of past events. 



But astronomers are not only able to predict they 

 can also trace back the paths of the celestial bodies, 

 and say : e At such and such a long-past epoch, a 

 given star or planet occupied such and such a position 

 upon the celestial sphere.' But how are they to 

 verify such a statement ? It is clear that, in general, 

 they cannot do so. Those who are able to appreciate 

 (or, better, to make use of) the predictions of astro- 

 nomy, will, indeed, very readily accord a full measure 

 of confidence to calculations of past events. They 

 know that astronomy is justly named the most exact 

 of the sciences, and they can see that there is nothing, 

 in the nature of things, to render retrospection more 

 difficult than prevision. But there are hundreds who 

 have no such experience of the exactness of modern 

 astronomical methods who have, on the contrary, a 

 vague notion that modern astronomy is merely the 

 successor of systems now exploded ; perhaps even 

 that it may one day have to make way in its turn 

 for new methods. And if all other men were willing 

 to accept the calculations of astronomers respecting 

 long-past events, astronomers themselves would be 

 less easily satisfied. Long experience has taught 

 them that the detection of error is the most fruitful 

 source of knowledge; therefore, wherever such a 



