228 Scientific Proceedings, Royal Dublin Society. 
Letts and Mr. Blake, every precaution is taken to ensure absolute 
accuracy and elimination of all errors. 
The absorbent is kept in a vessel coated inside with paraffin wax, 
and so arranged that, when a charge is withdrawn, its place is taken- 
by air freed from its carbonic anhydride. The absorbent, therefore, 
retains its strength unchanged as I have verified for a period of forty 
days. (Fifty cubic centimetres of baryta water required 48°32 cubic 
centimetres of standard hydrochloric acid, one cubic centimetre of 
which was equal to 1 cubic centimetre of carbonic anhydride. After a 
lapse of forty days, 50 cubic centimetres of the same solution of baryta 
water required 48:14 cubic centimetres of the above standard hydro- 
chloric acid, an almost inappreciable alteration.) 
The charge of the absorbent is transferred to the receiver without 
coming in contact with the atmosphere. The titrations are performed 
im vacuo, with very delicate burettes, and, as has been shown, with a 
more delicate indicator than turmeric paper; and, lastly, as Professor 
Letts and Mr. Blake have shown, the error arising from the action of 
the absorbent on the glass of the receiver is also eliminated. 
In Pettenkofer’s method of manipulation, there are no precautions 
taken to ensure the constancy of the strength of the absorbent, nor 
care taken to avoid errors due to absorption of carbonic anhydride 
from the air during the transfer of the absorbent to the vessel in which 
the titrations are made and during the titrations themselves. It is 
not to be wondered at, then, that errors arise; but it is somewhat 
surprising to find that they may be so large. 
It may, however, be remarked that, as I have already mentioned, 
various modifications of Pettenkofer’s process have been employed in 
which these sources of error are to some extent avoided. Thus F. 
Schulze! employed turmeric tincture as indicator, and, when titrat- 
ing, pierced the rubber sheet tied over the absorbing vessel with the 
end of the burette. 
To confirm the foregoing results, known and definite volumes of 
carbonic anhydride, at definite temperature and pressure, were added 
to the large flasks of about six litres capacity, and titrations performed 
according to Pettenkofer’s method, as far as possible, at least. The 
absorbent had to be added and withdrawn without coming in contact 
with the air; but the subsequent manipulation was performed as in the 
Pettenkofer process. Thus only about half the process was really 
Pettenkofer’s. Simultaneous experiments were performed according 
to the method of Professor Letts and Mr. Blake. 
1 Landw. Vers. Stat., 9 [1867], p. 217. 
