THE MOLECULAR VOLUMES OF LIQUIDS 193 



cessors, since it is for the most part the result of comparatively 

 recent attempts to unravel molecular structure. Although 

 they, no doubt, fully recognised that important clues concern- 

 ing the connection between molecular volume and chemical 

 composition might be expected to follow from the study of the 

 liquid hydrocarbons, their want of knowledge as to the essential 

 differences in molecular arrangement which may occur among 

 the members of this large group led them to compare substances 

 between which no true analogy exists, and from which no 

 valid or rational deductions could be drawn. Progress under 

 such conditions was therefore impossible. And this leads to 

 the general remark that anything in the nature of sound 

 comprehensive generalisations concerning the connection of 

 the physical quantity we connote by the term molecular 

 volume and chemical nature, using that phrase in its widest 

 sense, can only be expected to follow when we are dealing 

 with groups of substances of which the structure and con- 

 stitution are understood, or which are at least known to be 

 related among themselves in a manner capable of precise 

 definition. Not that this should discourage the accumulation 

 of the necessary experimental material even in the absence 

 of such knowledge. We agree with Mr. Le Bas that a know- 

 ledge of this particular physical constant should be regarded 

 as necessary to the complete history of a chemical compound, 

 even although under our present limitations we may be unable 

 to interpret its full significance. As it is, experience has 

 shown that this constant has served to afford a decisive clue 

 to constitution when chemical considerations alone have given 

 equivocal or contradictory results. 



Although the actual numerical values of the molecular 

 volumes given by Kopp for the particular hydrocarbons he 

 studied have been found by subsequent observations to be 

 substantially correct, proving the high degree of accuracy of 

 his experimental work, the inferences he drew as to the funda- 

 mental values to be assigned to the respective atoms of carbon 

 and of hydrogen are wholly invalid for the reason already 

 given, that he drew no distinction between hydrocarbons of 

 essentially different constitution ; hydrocarbons of open-chain 

 structure, for example, being treated as strictly comparable 

 with those of ring structure. Further investigation has 

 established that Kopp's assumption that similar atoms have 



