N bibs on Arctic Air, 39 
out by Thorpe—largely qualified by the methods and mode of 
manipulation adopted. 
The estimations were made by Pettenkofer’s method, baryta 
being used as the absorbent, and decinormal hydrochloric acid for 
the titrations. The latter were made in the presence of the pre- 
cipitate, and with the whole of the baryta solution used, together 
with the rinsings of the bottle. In ascertaining the point of 
neutralization it was found necessary to discard litmus, since 
by candlelight it somewhat masked the disappearance of the tur- 
meric ring. 
In ordinary temperatures the process is perfectly simple, but a 
little difficulty was experienced in applying it in very low tem- 
peratures, because the samples of air had to have their tempera- 
ture raised one hundred degrees Fahrenheit before they could be 
subjected to the action of the baryta solution in the temperature 
of my cabin. It was therefore necessary to make arrangements 
which would admit of the escape of the expanding air without 
allowing admixture with the air of the ship. 
In my first experiment 4735° of air were taken on 10th 
December from.two feet over the floe; one hundred yards to 
windward of the ship ina N.N.W. wind ; force 3 and temperature 
—14'8 degrees (F), pressure 29°52. The expansion was arranged 
for by opening the bottle under mercury, and leaving it inverted 
with its lips touching the mercurial surface till the escape of 
bubbles had ceased. The results corrected for temperature of 
pressure showed a percentage of -06417 of carbonic acid. 
This plan was not satisfactory, because it required close atten- 
tion to prevent regurgitation of air into the bottle from slight 
changes of temperature in my cabin, and the air in the bottle had 
to be exposed for a moment to contamination while the baryta 
was being introduced. 
For the next experiment the stopper of the bottle was replaced 
_ byavarnished cork perforated by two glass tubes with stop cocks; 
the one to admit the air and baryta solution; the other to 
‘ allow of expansion through a mercurial valve. But this arrange- 
ment interfered with the shaking of the bottle. | With it on the 
18th January I examined 9565 of air at a temperature of —40 
degrees, and obtained a percentage of ‘0485. 
A final improvement in the apparatus was made by substitut- 
