( 221 ) 

 SERIES III. 



From this follows that as the equilibrium of pressure and tempe- 

 rature is obtained better, the densities of the phases become more 

 and more equal, and that at last after the application of the tempe- 

 rature correction only very small differences remain. 



The much more considerable deviations found by de Heen 

 (o//o t . = 1.19, see Coram. No. 68) must, therefore, be attributed to 

 admixtures and differences of temperature. ') '). 



§ 4. Example of differences of density as found by de Heen, 

 caused by a slight impurity of the CO,. 



The great influence of small quantities of admixture is very con- 

 vincingly shown by the following results. 



In a group of determinations ending with a repetition of experi- 

 ment 2 Series I we found : 



It appeared that in consequence of carelessness in the cleaning 

 after it had been repaired, in the metal bottle in which the pure 



l ) That during the opening of the cock no important exchange of liquidogeneous 

 and gasogeneous molecules between the upper and the lower reservoir can have 

 taken place (to use the terminology of de Heen), appears from this, that when the 

 upper reservoir was filled with air and the lower reservoir with air with 31 pCt. 

 GOo, and the equilibrium was established without the cock being opened, after 

 the cock had been opened for 5 minutes, only 0,33 percent of C0 3 had passed into 

 the upper reservoir. 



3 ) They give both systematic errors of the same character (cf. § 1 p. 2 note 1 

 above). 



