ERRORS OF DATA AND OF THE SENSES. 115 



last longer than the phenomena themselves ; for instance, 

 a rapid succession of sparks appear like a continuous 

 light. A mirror proves the deceptive nature of vision 

 when uncorrected by touch or by knowledge. In every 

 single act of observation also there may exist very nume- 

 rous causes of error ; and non-observation of them does 

 not by any means prove their non-existence. It often 

 requires a high degree of scientific knowledge and train- 

 ing to expel the errors of the senses. 



When prejudiced persons, or those who are undis- 

 ciplined in correct observation, observe phenomena, they 

 are very apt to notice and remember the circumstances 

 which are favourable to their mental tendencies, and 

 neglect or forget the unfavourable ones; and this is a 

 common source of a whole multitude of popular errors and 

 delusions which hinder the progress of scientific truth. 

 We should therefore always reject or weigh carefully the 

 observations and conclusions of an incompetent, biassed, 

 or prejudiced person. Lord Bacon says, ' Men mark when 

 they hit, and never mark when they miss.' He quotes 

 also an ancient story of a man who was shown a temple 

 containing portraits of a number of persons who had paid 

 their vows before going to sea, and had not been drowned ; 

 and when asked to acknowledge the power of the gods to 

 preserve persons from shipwreck, replied, ' Aye, but where 

 are they painted who were drowned after their vows ? ' 



Personal bias or antipathy, or a state of expectation, is 

 very apt to influence us without our perceiving it ; and in 

 order to exclude this in chemical analysis, we make what 

 is called 'a blind experiment,' that is, we analyse a 

 weighed quantity of the substance, the weight of which is 

 unknown to us until after the analysis is complete ; we 

 then compare the total weight of what we have found with 

 the weight of what we had taken. With some scientific 



i 2 



