248 ANNUAL EEPOET SMITHSONIAN INSTITUTION, 1908. 



value that, inversely, from the appearance of these properties we can 

 predicate the existence of these definite forms of combination. The 

 classic example of this type is afforded by the power of optical rota- 

 tion in carbon compounds, which is dependent on the existence of 

 one or more asymmetric carbon atoms (or similar asymmetric struc- 

 tures) in the molecule. 



In the same way in these organic compounds one can discover from 

 the appearance of color, or more exactly, from the appearance of 

 certain characteristic absorption bands, as well as of fluorescence 

 the existence of particular groupings in the molecule. To this same 

 category of properties belongs also, in the largest sense, electrolytic 

 conductivity, which indicates the existence of free ions — that is to say, 

 the combination of elements or radicles with electrons; and lastly, 

 the appearance of the maximum value of 5 :3 for the ratio of the spe- 

 cific heats of a gas is, according to the kinetic theory of gases, an 

 indication of a monoatomic state, a conclusion which, as is generally 

 known, was first applied to mercury vapor, and which in recent times 

 has been of inestimable service in the determination of the atomic 

 weights of the argon group of elements. 



In the last analysis all these properties are probably constitutive 

 also, and their interpretation as either purely molar or purely additive 

 is only a more or less close approximation. Often, even in the regions 

 which have already been cleared up to a great extent by the additive 

 method, great difficulties have appeared upon a more careful investi- 

 gation. I should like, therefore, all the more to place before you a 

 special case where the theory seems to have reached the greatest 

 exactitude. 



We have succeeded in reducing to exact terms the gas densities 

 which for a long time afforded the only means of molecular weight 

 determination. On account of the variation from the laws of per- 

 fect gases, which all actual gases show, it was natural that a method 

 of approximation should be developed. In our epoch, however, we 

 have learned with the help of van der Waals' equation, particularly 

 by using compressibility, to reduce all gases to the ideal gaseous 

 state and thence to deduce the exact relative values of molecular 

 weights. 



From these results, two ends have been attained : First, it has been 

 proved that the most important of the theoretical laws that we 

 possess, Avogadro's rule, appears to be an infinitely exact natural law ; 

 and second, a new method, purely physical, of determining atomic 

 weights has been acquired which can stand comparison in exactness 

 with the methods of analytical chemistry, but which is limited natu- 

 rally to the cases where the density and compressibility of a chem- 

 ically pure gas can be determined. 



