426 PROCEEDINGS OF THE AMERICAN ACADEMY § 18 



of these eleven liquids, whether we reject certain of them 

 or not; it is equally obvious that our own formula here 

 affords a still closer approximation. In the case of the sev- 

 enty-five liquids of Table VII., it will be found that the 

 average ratio between the coefficients, A^ B and C, corres- 

 ponds a little more closely to the Theory than to the em- 

 pirical formula, being nearly midway between the two. 



There are not sufficient data to prove that either formu- 

 la may not be absolutely true; but the indication of such 



