1G4 
TIIE VOYAGE OF H.M.S. CHALLENGER. 
In this manner 7 experiments were carried out at the temperatures 1 6° '5, 17°'3, 17°‘4, 
1 n '0, 20 0, 21 0, 25 c- 2 C., and a curve was drawn through points indicating the values 
■ ;3 r Tin- v suits were less irregular than those of the preceding series, but again they 
w. iv higher than Bunsen’s. In the last two experiments the saturation of the water was 
• in the boiling-out flask itself, to avoid transference from one vessel to another, 
:h. fla-k having been provided for this purpose with a kind of egg-shaped “allonge,” 
attaehed to its neck by an india-rubber tube. Before quoting the results, I will state 
that this “ nitrogen series” was followed by a new series of eleven experiments with air 
from carbonic acid at the temperatures 3 0, 3, 4 0, 5,4 0, 7, 1 6° *7 , 17°'0, 1 7°*8, 18° - 5, 
20°‘2, 20 r- 2, 2G°1, 30° - 0 C. The results again were not up to my expectations; 
what they were will be stated presently. I prefer first to report on a lengthy series of 
at; -mpts to d< t ermine the coefficient fix for oxygen directly by experiments with pure 
-xyg- n gas. In all these experiments with oxygen the saturation was effected in the 
b"iling-out flask to avoid absorptiometric exchange with the atmosphere. The oxygen 
xira ud from the liquid in which it had been absorbed was always tested, after having 
1 ■ u measured, with alkaline pyrogallate, when it invariably disappeared, leaving no 
measurable residue. 
The first eleven experiments were made at the temperatures 5°’l, 5° - 3, 13°‘2, 13 0, G, 
14°6, 15°‘5, 1 5 0 o , 18°'G, 21° - 8, 21°’9, 23 '7, 28°'G C. The results again were not as 
r gular as 1 should have wished, and the values /3 l were invariably less than the corre- 
sj" mding ones calculated from the values of X of the series of experiments upon air. 
Wh nee I concluded that some of the oxygen had been absorbed by the india-rubber 
• .pp r "f tie' .laeobsen flask in the boiling-out process. I accordingly caused Mr. Lennox 
t > oiistruct a boiling-out flask entirely of glass, and after numerous failures he succeeded 
in producing one which worked not unsatisfactorily. The results which we obtained 
with tin- apparatus were not any higher than the previous ones, and even less constant. 
I tier* fore do not describe it here any further than by saying that the india-rubber 
-topper wa- replaced by a hollow glass stopper, to the upper end of which the pear- 
-haped bull* was fused on. The stopper had a small perforation about half-way 
b* tween the top and the bottom end of its working surface, which, when the stopper 
v. is turned into a certain position, just met the upper end of a groove in the ground 
exl nd< i downwards to the end of the ground part. It will readily be 
that tin- arrangement was equivalent to the lateral hole in Jacobsen’s gas- 
bulb -tern in theory; in practice, it did not work at all satisfactorily, because very 
often the stopper would Stick so fast that it was impossible to turn it round. On more 
! ;■ ' u< h occasion the apparatus broke and had to be renewed. Of the many 
ii to work it only seven were carried through without accident; and even these 
I look uj*on with suspicion, because of the numbers for /3 l which they gave, five were 
1 ‘ 1 ould have been expected from the sum-total of previous corresponding 
