( 548 ) 



— 10.9°, even when the observable mehing range only amounted to 

 some tenths of degrees. In these determinations, which were made 

 in a bath of carbonic acid and alcohol, which could be kept constant 

 for a long time by insertion of bits of carbonic acid, the tubes were 

 heated very slowly, while they were continually shaken ; the tempe- 

 rature was read on an Anscliiitzthermometer, the mercury thread 

 of which was quite immersed in the bath. 



If the heating took place at the rate of about 0.1° per minute, a 

 melting point which was appreciably too high was found in spite of 

 the shaking of the tubes. Whether the melting-point of the N^O^ can 

 depend on the circumstances to which it was subjected previously, 

 and the deviations can possibly be explained by the formation of 

 mixed crystals of N,0^ and NO,, will have to be decided by a 

 further investigation by Prof. Smits and one of us. 



§ 7. The nitrogen tetroxide may justly be called the classical 

 example of dissociation. Numerous scientists have been engaged with 

 the experimental determination and specially with the theoretical 

 considerations and calculations. As basis of these calculations serve 

 the determinations of the vapour density of dissociating N^O^ under 

 varying temperature and pressure. From these determinations among 

 which those of the two brothers Natanson^) are the most accurate, 

 ditferent theorists have derived several expressions for the constant 

 of dissociation as function of the temperature. The great variety in 

 the proposed formulae is owing for the greater part to the different 

 values ascribed to the specific heats of the NO, and N^O^. As these 

 are not sufficiently known in their dependence on the temperature, 

 it is only possible to determine this relation roughly. 



Among the calculations carried out with the greatest accuracy, 

 those by Schreber*) deserve a foremost place. 



With the aid of his recalculated expression :. 



4 a' F 2866.2 

 log — = 9.13242 , 



which represents the observed densities wnth sufTicient accuracy the 

 degree of dissociation can, accordingly, be calculated for given 

 temperature and pressure. , .^ . ' . '• v' 



We have carried out these calculations for the vatues of the 

 pressure and the temperature along the P-T line to get an impression 

 of the change of the degree of splitting up of the N^O^ along the 



^) Wied. Aon. 24 454 (1885Ï; 27 606 (1886). 't r,- v ;H 



2. Z. ph. Gh. 24 651 (1897). '.) 



