378 FEAGMENTS OF SCIENCE. 



ment. The quantity of these latter possessed by the 

 first pair, is not simply diffused among their descend- 

 ants, for in that case the last must feel more feebly than 

 the first. The sensations and feelings must necessarily 

 be referred back to the flour, where they exist, weak and 

 pale it is true, and not concentrated as they are in the 

 brain.' 1 We may not be able to taste or smell alcohol 

 in a tub of fermented cherries, but by distillation we 

 obtain from them concentrated Kirschvvasser. Hence 

 Ueberweg's comparison of the brain to a still, which 

 concentrates the sensation and feeling, pre-existing, but 

 diluted in the food. 



'Definitions,' says Mr. Holyoake, 2 'grow as the 

 horizon of experience expands. They are not inventions, 

 but descriptions of the state of a question. No man 

 sees all through a discovery at once.' Thus Descartes's 

 notion of matter, and his explanation of motion, would 

 be put aside as trivial by a physiologist or a crystallo- 

 grapher of the present day. They are not descriptions 

 of the state of the question. And yet a desire some- 

 times shows itself in distinguished quarters to bind us 

 down to conceptions which passed muster in the infancy 

 of knowledge, but which are wholly incompatible with 

 our present enlightenment. Mr. Martineau, I think, 

 errs when- he seeks to hold me to views enunciated by 

 4 Democritus and the mathematicians.' That definitions 

 should change as knowledge advances is in accordance 

 both with sound sense and scientific practice. When, 

 for example, the undulatory theory was started, it was 

 not imagined that the vibrations of light could be trans- 

 verse to the direction of propagation. The example of 

 sound was at hand, which was a case of longitudinal 



1 Letter to Lange : ' Geschichte des Materialismus,' zweite Aufl, 

 vol. ii, p. 521. 



Nineteenth Century,' September 1878. 



