Carbon-Dioxide Content of Sea-Water at Tortugas. 91 



to establish the relationship definitely. Even wider differences are 

 shown in Dole's figures on different days. No diurnal variations were 

 found in the excess base, and it is doubtful whether any will be, as solu- 

 tion and precipitation of carbonates from sea-water are not thought 

 to be an action that reaches equilibrium rapidl}''. Even the final CO2 

 distribution attained between sea-water and air appears to be reached 

 slowly, according to the experimental evidence at present available. 



Determinations of CO2 should probably be made soon after the 

 time the samples are collected, on account of the possibility of the 

 decay of organic matter, such as algae, in preserved samples. More- 

 over, the writer^ has shown that sea-water preserved for a long time 

 in glass loses part of its CO2, owing to the precipitation of aragonite 

 brought about by the slow solution of alkali from the glass. This was 

 ordinary glass. It is well known that " nonsol " glass is attacked more 

 slowly. The other change found accompanying the decrease in CO2 

 was a decrease in the titration alkalinity of the water, corresponding 

 to the decreased quantities of carbonate and bicarbonate, but the 

 Ph value was not greatly changed, as sodium had merely taken the 

 place of calcium. 



The average "excess base" found at Tortugas corresponds to a 

 normality of 0.00239. This titration includes everything that con- 

 sumes acid; it represents chiefly bicarbonate, about 0.00183, some 

 carbonate, about 0.00041, and other substances that contribute to 

 the alkalinity, about 0.00015. The last figure, however, was not deter- 

 mined at Tortugas, but with Gulf Stream water that had been shipped 

 to Washington. This water was collected November 30, 1919, tem- 

 perature 17.94° C, lat. 24° 24' 20" north, long. 81° 31' 15" west, 

 about 7.5 miles from American Shoal Light, depth 105 meters. Its 

 titration alkalinity, or excess base, was 0.00258, total CO2 0.0968 

 gram per liter, Ph 8.08 at 25° C. These figures give on calculation 

 bicarbonate alkalinity 0.00197, carbonate alkalinity 0.00046, other 

 alkalinity 0.00015. The observations here reported are too few, 

 however, to warrant discussion. 



The methods used in arriving at these figures are given below. 



METHODS. 



Carbon dioxide. — The total CO2 was determined by adding an 

 excess of hydrochloric acid to 500 c. c. portions of the water and boiling 

 about 15 minutes, while a current of CO2 free air was passed through 

 the water, then over calcium chloride, and finally through two soda- 

 lime tubes, each having its last third part filled with calcium chloride. 

 The soda-lime tubes were weighed with a counterpoise to minimize 

 errors likely to be caused by the high humidity. The second tube 

 served as a check on the absorbing power of the first. 



1 Jour. Wash. Acad. Sci., 10, 249 (1920). 



