Physics and Chemistry. 55 
carbonates. In the autumn of 1878 the experiments of Tornde 
were again resumed, after an interval of about a year. 
use of an improved ‘form of apparatus devised by Classen, the 
total amount of carbonic acid yielded by sea-water on boiling 
was determined, and found to agree very closely with that 
obtained by Jacobsen’s method. The determinations were, how- 
ever, carried still further, and not only the total amount measured 
ates, and the difference between the two gave the amount com- 
bined with bicarbonates, for it is shown that the onion ObaGne va- 
tions can only be e xplained in case the water contains no free gas. 
Tornée gives a table in which the results of sey “eo aene determin- 
ations, made with water taken between 65° and . latitude, 
rom various depths, from the surface down to Hee fathoms. 
The table shows a remarkable constancy in results, the amount of 
Ww 
points near the ice on the coast of Greenland are excluded, the 
extreme ja dictaao are only 4 mgr. and 8 mgr. respectively. 
average result, the amount of carbonic acid forming neu- 
tral carbonates i is be ate o be 
52°78 + 0°083 mer. 
per liter, with a probable error sor 3 oe observation of 0°662 
megr., and for that forming bicarbon 
43°64+ 0°16 mgr. 
with a probable error of 1°26 mgrs. per liter. A series of re ters 
ments were also carried on to explain the liberation of gas 
the carbonates on boiling, and the conclusion is reached Ghat: this 
is due to the slow te action of the carbonates and salts o 
magnesia it contai In regard to the character of the carbon- 
ates present in sea-w ribek Tornde believes it probable that be 2 are 
rather sodium and potassium carbonates than carbonate of cal- 
cium, this supposition ser ats toe explain the alkaline reaction of 
ivi noted by him, and also the resistance to decomposition ob- 
served, allowing of hciy being exposed for an hour ether" toa 
temperature of near rly 90° C. without effect. Earlier _determins- 
e 
not witb as to allow of even appr oxim ste pauses. 
n the Action of the Oxides of Nitrogen on Glass at a high 
teinperatere ; by Ett ai MorGan.—Several attempts were 
made to determine the trogen in organic nitro-derivatives by 
heating the substance in aslo d tubes of hard shee containing 
oxygen, and in presence also of a little mercury to effect the 
reduction of any oxides of nitrogen which might be $onndee. but 
in all cases the nitrogen obtained was dots deeABIS less than the 
