104 
taken up; and such a brine is useful in many cases in apply- 
ing this test. or preservation of liquids under test from the 
air, use was made of petroleum (previously freed from 
oxygen by sodium), and of different hydrocarbons, including 
in case of hot liquids, paraffine, which, on cooling, forms for 
a time a solid seal: fixed oils were also used from the first. 
Such liquid seals were found but transiently to protect 
against the 8 Ibs. per square inch pressure of aerial oxygen. 
All the oils in time admitted it, either chemically or mechan- 
ically. . 
To collect and transport samples of natural waters, I draw 
down small at points a tolerably heavy glass tube, fill with 
hydrogen or nitrogen (first washed with alkaline pyrogalline), 
and seal at these points. One of the elongated bulbs thus 
made is introduced into the bottle of natural water to be pre- 
served, one point broken off under the water, the bottle then 
hermetically stoppered or sealed when completely full. Hx- 
pansion will not then break it. It must be opened only under 
quicksilver (previously wetted with alcohol to detach air), 
and portions of the water passed into a eudiometer for analy- 
sis. Carbonic acid gas upon the surface, was found preferable 
to liquid hydrocarbons, in protecting from air. Causing, for 
this purpose, the CO’ to be evolved from the liquid itself, a 
most important discovery was made, viz.: that free oxygen 
cannot coéxist in water with an excess of CO’, within ordinary 
atmospheric pressures and temperatures. This new principle 
has numerous applications in almost every science. 
All effervescent waters, natural and artificial, are found 
free from oxygen. ‘Plain soda water,” is a very convenient 
menstruum for these experiments. Beer and champagne: 
also contain no oxygen, proving the principle to hold also at 
higher pressures. Quantitatively, I have so far generally 
used, for comparison, a solution of permanganate whose 
coloring power with pyrogalline and ammonia has_ been 
normalized to some standard. The value of the color and 
shade must of course be referred at the start, and as often as 
practicable, to that given by actual aerated water whose 
oxygen has been measured directly, as usual, by boiling and 
exhaustion. Beside permanganate, other oxygen solutions 
have been tried, with more or less success. 
I apply pyrogalline also in another way altogether new— 
to determine the rates of absorption of aerial oxygen by 
different waters and other liquids. Like columns of different 
natural waters, ceteris paribus, exhibit progressive penetrations 
of air to greatly varying depths in like times. 
