170 
THE VOYAGE OF H.M.S. CHALLENGER. 
Then is no need for any further description ; the principle of the apparatus is sufficiently 
explained ; and whosoever may reproduce and use the apparatus will, of course, have 
t ■ acquire a certain degree of familiarity with the apparatus before he can work it properly. 
T<» determine a value X, about 1 to 1^ litres of water or sea-water were placed in a 
•Winchester quart” bottle, standing in a large water-bath of rigorously-constant 
t, nip' rat ure, and exposed for several hours to a current of air free from carbonic acid sucked 
in from the outside open atmosphere, first through a large soda-lime tower, and then through 
tie water bv means of a filter-pump. Very frequently the bottle was lifted out of its bath 
for a few seconds, violently shaken, and immediately replaced in the bath. The 
t nip- lature of tin* water operated upon was always taken directly by means of a delicate 
thermometer plunged into it ; the thermometer in the bath merely served to control the 
temperature there with regard to its constancy. When absorptiometric equilibrium 
was supposed to have been attained, the water, by means of a wide and long-necked 
funnel, was transferred to the Jacobsen flask, a considerable quantity being allowed to 
flow over t<> eliminate the influence of the outer air as much as possible, the perforated 
india-rubber stopper inserted, and the stem of the “gas-tube” inserted so that the lateral 
hole was shut up by the mass of the india-rubber. The pump was then worked until the 
_* is-tube was completely exhausted. The gas-tube stem was then pressed down so as to 
'ring the lateral hole within the water, and the water heated to boiling while the pump 
was b.-ing wrought at short intervals to maintain a sufficient vacuum, which at the end 
v : rais'd to the highest attainable pitch. The boiling was continued until after the 
- \li ust-bulb of the pump was emptied for the last time, five minutes’ further boiling 
failed to extract a visible gas bell. The gas was then measured and analysed. 
We began with a long series of trials with distilled water, and then passed on to experi- 
ment with sea-water. The latter was prepared synthetically on a large scale from pure 
-dts nd distilled water, so as to represent about an average Challenger water. Care was 
t. li.en to give to the solutuin very nearly the quantitative composition demanded by my 
"wn hi dv- -, as reported in the first part of this memoir. The “alkalinity” was estab- 
by addition of the calculated proportion of ignited carbonate of soda to a neutral, 
but otherwi.-e < orreetlv adjusted, water. The following is a statement of the results. 
Experiments upon Pure Water. 
-7 ■ vj„ lament* were made at temperatures varying from 0°’5 to 47° C. The values 
• 1 g i w< re plotted and united by the nearest curve, which enabled me to 
■ p'Midt* which were obviously infected with bad unobserved errors, and to 
A\ 
b ' enough valuer for the expression ^ f to reduce any group of values obtained 
