ATOMIC WEIGHTS 49 



Corrected, by reduction to latitude 45°, etc., these two series become 

 nearly identical with each other, and with Guye and Pintza's average; 

 namely, 0.77085 and 0.77086; in mean, as one series, 0.770855, ±.000034. 

 With Guye and Pintza's figure, the general mean becomes 0.77083, ± 

 .000024. Hence, with Morley's weight for a litre of oxygen, the crude 

 density ratio is 



O.tNH,,: : 32: 17.2619, ± .00063 



The law of Avogadro, that equal volumes of gases contain equal num- 

 bers of molecules, is rigorously true only for ideally perfect gases. Por 

 gases as they actually occur it is approximately true, but with varying 

 degrees of divergence. The approximation is close for the so-called per- 

 manent gases, while those wliich are easily liquefiable conform less nearly 

 to the law. In order, therefore, to compute molecular weights from 

 observed gaseous densities, it is necessary to apply corrections to the 

 experimental data, or else to employ methods of determination of great 

 manipulative difficulty. By measuring densities at very low pressures, 

 quite close approximations to the truth may be obtained, and observations 

 at high temperatures are also nearly valid. For example, Eayleigh ^ 

 from gaseous densities at very small pressures, obtained the following 

 value for nitrogen, as compared with the standard, oxvgen: 

 N, = 28.018. and N = 14.009 



On the other hand, by measin-ing the density of nitrogen at 1067.4°, 

 Jaquerod and Perrot" found 



N, = 28.0155, and N = 14.0077 



These values are probably not far from the truth, and are obviously 

 well in accord. At low pressures and at high temperatures gases are more 

 nearly in agreement with Avogadro's law than they are under ordinary 

 conditions. 



In the case of the oxygen-hydrogen ratio, the density corrections were 

 determined by actual measurement of the volumes in which the two 

 gases combined, a method which is not always applicable, or at least not 

 conveniently so. It is easier to compute the corrections from pliysical 

 data, and for this purpose various methods have been proposed.^ 



The following formulge, based upon the celebrated gas equation of Van 

 der Waals, are, according to Guye,* available for the reduction of gaseous 

 densities to true molecular weights : 



^Proc. Roy. Soc, 73, 153. 1904. 



^Compt. Rend., 140, 1542. 1905. 



3 See D. Berthelot, Journ. Physique (3), S, 263. 1899. Lediic, Ann. Chini. Phys. (7), 15, 1. 1898. 

 Guye and Friderich, Arch. Sci. Phys. Nat. (4), 9, 505, and 13, 559. Guye, Journ. Chim. Phys., 

 3, 321, and 5, 203, also Compt. Rend., 138, 1213, and 140, 241. There is a copious literature upon 

 this subject. 



* Journ. Chim. Phys., 3, 321. 1905. 



