REMARKABLE INSTANCE OF ERROR. 109 



paper, though, had he been aware of the consequences to 

 their full extent, I am persuaded that he would not have 

 done so. It comes to be a question whether, had Dr. 

 Wollaston informed him of the whole secret, Mr. Chenevix 

 would have been convinced.' L An instructive moral may 

 be drawn by a scientific investigator from this example, 

 especially the great danger of being too strongly impressed 

 with a preconceived idea, and the duty of not holding an 

 hypothesis as if it were a fixed truth. Nothing, also, so 

 effectually destroys the motives for research and the 

 pleasure of such occupation, as to find, after having made 

 and published a laborious investigation, that the conclu- 

 sion was all a mistake. 



It is important for a young investigator to be aware 

 of the possible extent of error. If we assume that to 

 know is to truthfully apprehend, we cannot know (al- 

 though we may believe) that which contradicts, or is 

 inconsistent with, natural truth i.e., we cannot truly 

 know an idea which is false in itself. In accordance with 

 this assumption, the region of possible error must be at 

 least co-equal with that of possible truth, because for each 

 truth there may exist in thought its negation, or opposite; 

 and the subject of mental error is complemental to that 

 of knowledge, because the two together constitute the 

 entire range of possible belief. I shall therefore confine 

 inyself to a consideration of the chief classes of errors 

 which are likely to be committed in making researches 

 in physics and chemistry. 



The great and primary sources of error are the im- 

 perfect action and limited extent of all our powers, and 

 especially that of the intellect. In order to avoid error 

 and arrive at truth, all our lower powers require to be 



1 History of Chemistry, vol. ii. p. 216. 



