122 



the same subject which have preceded it, on many points it is in error ; and he who 

 assumed it to be all perfectly correct would certainly be in the wrong. 



" It is owing to this circumstance that we can have no such thing as a standard 

 work on any branch of science. If we take up a scientific book, written twenty years 

 ago, on any subject which has been much worked at subsequently, we cannot help 

 being struck with the vast amount of error which it contains, — error sometimes on 

 points of great scientific importance. 



" Is there any reason to doubt that twenty years hence similar faults will not be 

 found in the scientific master-pieces of the present day ? 



" I will instance a case in point, by way of making myself perfectly intelligible. 

 In Duponchel's ' Lepidopteres de France,' Vol. vii. Part 1, p. 396, we find the asser- 

 tion that Cucullia Scrophulariae is only a pale and small variety of C. Verbasci. ' It 

 is sufiicient to breed a certain number of larvae of this latter in order to obtain this 

 variety ; M. Marchand pere, a very good observer, has several times had experience of 

 this.' When, some years ago, I was going systematically through Duponchel's work, 

 I was much struck with this statement ; for in this country the species were always 

 reputed distinct, though few could point out wherein the distinction lay ; and, as Du- 

 ponchel's statement rested on the repeated observations of ' a good observer,' the infe- 

 rence would naturally be that the statement would be correct. But, turning to Du- 

 ponchel's ' Supplement,' Vol. iii. p. 412, published in 1836 (nine years after Vol. vii. 

 Part I), we find ' We were wrong in supposing, on the testimony of M. Marchand, 

 that Cucullia Scrophularise was only a variety of C. Verbasci ; it is a good distinct 

 species, which in truth much resembles the latter in the perfect state, but of which the 

 larva is very diff"erent; ' and as, of course, though the correction refers to the previous 

 error, the error has no means of referring to the subsequent correction, those who read 

 the first statement, assuming all they find in that volume to be true, would be neces- 

 sarily led into error, from which it may be years before they get freed. 



" Those writers who have studied to most effect the works of previous authors na- 

 turally become very careful not to state too dogmatically any circumstance concerning 

 which they have the slightest lurking doubt; but it will sometimes happen that other 

 facts, which they had considered fairly established, and of which they had in their own 

 minds no doubt, are still subsequently found to be untrue ; and hence it is that the 

 very best scientific work roust always be studied, with a scrupulous care lest we be led 

 by it to place implicit faith on statements which are not true. 



" It frequently happens that a writer on another branch of science will make use of 

 some statement that he has met with in a scientific work, and base on it some long 

 argument and ingenious theory, which, where the original ground-work is found to be 

 untrue, becomes simply ridiculous. 



" This caution in the use of scientific works is necessary to the most advanced of 

 us ; but it is especially needed by the young student. He naturally looks up with 

 reverence and respect to the great names of the Professors of Science, and, unless 

 specially warned of the danger, would think it a sort of treasonable scepticism to 

 doubt anything which has been written by authors of such celebrity ; but the time 

 m;i7Z come (if he continues to s^Mrf?/) when he will find that his confidence and reli- 

 ance had been misplaced ; and then he will be inclined to atone for his former idolatry 

 of his favourite authors, by an unjust and harsh criticism of them, because the subse- 

 quent experience of men of science had proved the previous writers in error on many 

 points. 



