Letts & Braxe—The Carbonic Anhydride of the Atmosphere. 225 
In the aboye series, with the Pettenkofer process, it was impossible 
to obtain results which differed from one another by less than 2. 
This was, however, overcome to a great extent in the third series of 
experiments by the use of very much larger flasks. 
The results of the third series of experiments are :— 
Tuirp Seri—EsS—CoMPARATIVE DETERMINATIONS. 
(Larer Recetvers.) 
The air samples for both processes were collected simultaneously. 
PETTENKOFER. Lerts anp Buaxr. 
CO, in CO, in 
Date. 10,000 parts} Duplicate. | Difference. || 10,000 parts | Duplicate. | Difference. 
of air. of air. 
I II. if II 
Feb. 28 4:03 3°98 05 2°95 2°96 01 
Mar. 1 3°94 3°95 01 2°90 2°90 00 
pin ied 3°96 3°92 "04 2°85 2°85 00 
Apftio 4:00 3°96 04 2°86 2°86 00 
we 1a 3°86 3°82 04 2°81 2°82 01 
Mean,.. 3°95 3°92 "03 2°87 2°87 004 
As a result of these experiments, we are thus confronted with the 
fact that Pettenkofer’s method of determination gives higher results— 
some 30 per cent.—than those obtained by the process devised by Pro- 
fessor Letts and Mr. Blake, and by other processes which are apparently 
reliable. Other investigators have also drawn attention to the high 
results obtained by Pettenkofer’s process. 
Kbermayer! obtained almost the same difference in comparative 
and simultaneous experiments with Pettenkofer’s ordinary process 
(‘‘ Flaschen methode’’) and the ‘‘ Aspiration methode.”’ 
It remained, therefore, to decide two important questions :— 
Is the Pettenkofer process at fault? and, if so, what is the 
reason ? 
1 Ebermayer, ‘‘ Die Beschaffenheit der Waldluft.’’ Stuttgart, [1885], p. 16. 
