226 Scientific Proceedings, Royal Dublin Society. 
To decide the first of these questions, several series of experiments 
were performed. In the first of these accurately measured volumes 
of carbonic anhydride at a given temperature and pressure were care- 
fully transferred to corked receivers containing only a small quantity of 
air, and nearly filled with a measured volume of absorbent. The 
apparatus employed for this purpose was similar to that figured on 
p. 130, and was used as there described—with the difference that the 
measured volume of carbonic anhydride was driven by pressure through 
a long tube which nearly touched the bottom of the receiver: the 
latter, however, being inverted. 
After absorption, the excess of the latter was determined precisely 
as in the Pettenkofer process. 
The results were as follows :-— 
Carsonic ANHYDRIDE aT N. T. P. 
Difference. 
Number. Taken. Found. 
Absolute. Percentage. 
| 
i °9185 c.cm. 1-210 c.cms. + °3015¢c.cm. | + 32 percent. 
2 EONS Oe ees 1:240 ,, + °3215 ,, + 34 30 
3 5942 Ola 12105, + :2680  ,, + 28 Dp 
4 "9420s, 1-230 ,, + :2880 ,, + 30 
by) ANGI) os 1:256 ,, th KO 4g + 30 AA 
6 s9O Os 1256 ,, + :2950  ,, + 30 50 
Mean, .. 31 per cent. 
These experiments indicate that the results obtained by Petten- 
kofer’s process are some 30 per cent. too high, thus agreeing with my 
previous experiments on atr. 
In order to ascertain the true source of error, a similar series 
of experiments was performed, but the absorbent was drawn off 
into the vacuous vessels, and the titrations performed as in Letts’ 
and Blake’s process, using phenol-phthalein as indicator, as it is 
obvious that it was impossible to employ turmeric paper under these 
conditions. 
