THE TEST OF TRUTH 



and results, and find no difficulty in imagin- 

 ing that some of the conditions varying might 

 affect the sum-total of results. Or if this also 

 be taken to imply too much conscious philo- 

 sophizing in uSj it is undeniable that our con- 

 ception of a crow, as of any other vertebrate, is 

 made up of a large number of conceptions, of 

 which the conception of blackness is not the 

 one upon which the specific identity of the sum- 

 total depends. We have had experience of bay 

 and of sorrel horses, of black and of white bears, 

 of gray and of tortoise-shell cats ; and, in ac- 

 cordance with such experience, we find it per- 

 fectly easy to regard any other animal as vary- 

 ing colour while retaining its specific identity. 

 Our belief that all crows are black rests, there- 

 fore, upon purely negative evidence, — upon 

 the absence of any experience of crows that are 

 not black ; and no amount of negative evidence 

 can outweigh a single well-established item of 

 positive evidence. 



Quite otherwise would it be if our explorer 

 should assert that he had discovered crows 

 destitute of a vertebrate skeleton. We should 

 reply, with confidence, that in the absence of 

 such a skeleton the animal in question could not 

 have been a crow. And the justice of the reply 

 becomes apparent when we turn to the case of 

 the nitrogen, where the conditions are so simple 

 that we can keep them all in mind at once, and 

 8i 



