456 SCIENTIFIC THOUGHT. 



because they follow from distinctly defined and inde- 

 pendent lines of reasoning. The three criticisms can be 

 summed up in three distinct arguments, all three de- 

 manding our special and exhaustive study. These three 

 arguments may be summarised as follows: 



First. The atomic view is a hypothesis resting upon the 

 fact that substances combine in fixed and fixed multiple 

 proportions, and upon the further observation that bodies 

 both in the solid and liquid state show different properties 

 in different directions of space. But as to the nature 

 of the differences of the elements the atomic view gives no 

 information; it simply asserts these differences, assumes 

 them as physical constants, and tries to describe them by 

 number and measurement. 



The atomic view is therefore at best only a provisional 

 basis, a convenient resting-place, 1 similar to that which 

 Newton found in physical astronomy, and on which has 

 been established the astronomical view of nature. 



Second. The atomic view in its present development 

 gives us no insight into the nature of those forces on which 

 depend the formation or destruction of chemical com- 

 pounds. It neglects the study of chemical affinity. This 

 must be conducted on different lines of observation and 



reasoning/ 



1 As these and other points re- 

 ferred to here will be taken up and 

 fully treated in future chapters of 

 this work, I abstain from giving 



Very suggestive in the first instance 

 is Lord Kelvin's address to the 

 mathematical and physical section 

 of the British Association in 1884, 



exhaustive references, limiting my- i reprinted in the first volume of his 



self to such writings as will ! ' Popular Lectures and Addresses,' 



give the reader a general idea of < p. 218, &c., "Steps towards a Kin- 



the various attempts which have ' etic Theory of Matter." 



been made to go beyond or behind j 2 In respect of this the Introduc- 



the Atomic View of Nature or to , tion to the first edition of Lothar 



supplement it by other views. > Meyer's ' Modern Theories in Chem- 



