337 



at the beginning of a determination with atmospheric air which, 

 after the diffusion is finished, is expelled through the three- 

 way-lap p and either collected in a recipient or directly taken 

 into the air-analysis-apparatus. 



For the air -analyses I applied during the Greenland- 

 expedition the extremely accurate and convenient Haldane- 

 apparatus with such modifications as were necessitated by the 

 special conditions. The apparatuses were carried for about two 

 months in an open skin-boat, they were often turned upside- 

 down and sometimes roughly handled, when we had to land 

 through the surf. Nevertheless, they did not get out of order 

 and always worked admirably. The gas-burette contained 10 cc, 

 and the part from 6.5 to 10 was graduated to O.oi cc. I read 

 it by means of a lens to 0.0005 cc, and the degree of accuracy 

 for the carbonic acid corresponded as a rule to this figure. 



Nearly all my analyses were made as double-determinations, 

 and I thus have a very good test as to their accuracy. I 

 give as an example the following quadruple-analysis of atmo- 

 spheric air^). 



-Vvn. 8 p.m. Ävütarpait. West-coast of Disko: 

 Volume taken 6.584 6.6065 9.971 9.960 



Absorption of CO2 .... 6.5805 6.603 9.9655 9.955 



Absorption of O2 7.874 7.866 



Percentage of COg .... 0.055 0.055 0.055 0.05 

 Percentage of O2 20.975 20.975 



The method described for equalizing the tension in a 

 sample of air with that of the water was only used during the 

 first part of the journey, when comparatively few determinations 

 were made. The apparatus was rather complicated and required 

 a comparatively large quantity of water, which had to be carried 



M As will be seen from the analyses here cited the percentage of carbonic 

 acid in the atmosphere was extremely high. This remarlcable pheno- 

 menon will be dealt with in the following paper. 

 XXVI. 22 



