41 



traces of carbonic acid are eliminated from such corks and no 

 difference was found in the double-determinations when one of 

 the tubes was analysed immediately and the other preserved 

 for a day or two. 



Errors may arise, during the analysis itself, from the 

 absorbing fluid, if this is not saturated with atmospheric air 

 exactly at the temperature and barometric pressure obtaining 

 (as mentioned in the preceding paper p. 346). When in Green- 

 land I did not know this source of error, but very often (though 

 not always) after a determination of carbonic acid I took the 

 sample of air once more into the absorption-pipette and satisfied 

 myself that no perceptible alteration of the volume took place. 

 By the numerous determinations which I have made since then 

 with more sensitive apparatuses I have never seen errors from 

 this source amounting to more than 0.002 "/o. It is rather pro- 

 bable that several of my Greenland-determinations are infected 

 with errors of this magnitude, because the temperature often 

 varied considerably. In almost all cases, however, the tempera- 

 tures at which the analyses were made were decidedly higher 

 than those to which the absorption-fluid was exposed during 

 the intervals. If anything, it would therefore be a little super- 

 saturated during the analyses and give off a little air. If my 

 analyses are infected by errors of this kind it must be admitted, 

 therefore, that they do not reach 0,005 ^/o and that they must, 

 in almost all cases, have had a tendency to diminish the figures 

 found for the carbonic acid. 



During my stay in Greenland I wanted to test the results 

 by means of another method, but my analytical outfit could 

 not be transformed into a PEiTENKOFER-apparatus or anything 

 Uke it. I was very glad therefore to find in the literature a 

 clear confirmation of my observations. During the Discovery- 

 Expedition Dr. E. Moss V) made 3 determinations of carbonic acid 



*) Notes on Arctic Air. Proc. Roy. Dubl. Soc. Vol. 2, 1880. 



