108 
TIIE VOYAGE OF H.M.S. CHALLENGER. 
in: i an ( ■•icuatcd flask of about l.l litres capacity, previously charged with the 
h.irvi . v itfi- in a manner precluding the premature formation of carbonate of baryta. 
Should a lilm of carbonate show in the flask, it is, of course, easy to empty it out (after 
1 i\::.g :ill"\vcd it to suck in pure air), to rinse it with water free from carbonic acid, and 
i hai-g.' and evacuate it once more. The air carrying with it the carbonic acid enters the 
flask th rough a glass stop-cock prolonged inside into a tube dipping into the baryta-water, 
so that the rate at which the gas flows in can be seen and regulated. In the rehearsals 
i - ' iir times occurred that the boiling liquid, through a momentary increase of pressure 
in the atmosphere of the flask was driven back through the air-inlet tube. This was 
easily prevented by the insertion of a mercury-valve between the flask and the air- 
gasom- tor. The air in the latter must, of course, be absolutely free from carbonic acid, 
or, if not, lie freed from it on its way to the boiling-flask. According to my experience, 
it is far easier to fulfil the former than the latter condition, by charging the gasometer 
with dilute caustic soda, and vigorously shaking the air with the liquid contents. When 
all the carbonic acid is in the vacuum flask . the latter is allowed to fill itself with pure 
air, and to stand until all the carbonic acid diffused throughout it may be assumed to 
1 m- absorbed. The glass stop-cock is then pulled out of its aperture in the india-rubber 
stoj.jM r, a few drops of turmeric solution are run in, the beak of the standard hydrochloric 
acid burette rted through the hole, and acid run into the liquid until it is almost but 
not quite neutralised. Should the point of neutrality be by chance overstepped, this is 
easily rectified by a few drops out of the baryta-burette. The hole is now closed, the flask 
moved about so aa to cause any trace of stray carbonic acid to be absorbed by the still 
decidedly alkaline liquid; the contents are poured out into a graduated cylinder, or, more 
< \v. nt ly, a t.-in-d phial, and weighed ; they are then allowed to settle in the absence 
- a diquot part is decanted off clear, and neutrality established by zig-zag titration 
with .i« id nd b.nyta, the point of incipient acidity being taken as the end-point. The 
in- ' ■ I :n this form, when applied to known weights of carbonate of soda as before, gave 
very satisfactory n suits. The apparatus employed is depicted in PI. I. 
A" : 1 iving thus come into possession of exact methods for the determination both of 
tv ■ ulM.nie ,-irid und the alkalinity, I next applied these in the following synthetical 
expert ments ; — 
I. Kxp< ri merits with Pure Solution of Bicarbonate of Soda. 
5 grma. <>f pure carbonate of soda, Ni 111 dissolved in water to 1 litre. 50c.c. 
item w re diluted with water to about 100 c.c., fully saturated with carbonic 
gas, and then diluted 2000 c.c. Of this solution successive volumes of 250 c.c. 
"‘h -ind ■ e h shaken n time with 10 times its volume of air, the air beino- 
r “ r * h -li iking. In the rc-ulting solution the total carbonic acid was deter- 
mined l*y the (“vacuum”) method. To check the synthesis the alkalinity of 
