48 RESPIRATORY EXCHANGE OF ANIMALS AND MAN 



I litre of distilled water saturated at 1 5 with atmospheric air (760 

 mm. dry pressure) (79 '04 per cent. N 2 , 20*93 P er cent. O 2 , and 0-03 

 per cent. CO 2 ) will contain according to Fox [1907, 1909] 



13*68 c.c. nitrogen 

 7 '22 ,, oxygen 

 0*30 ,, carbon dioxide 



corresponding to the absorption coefficients 0*0173 for nitrogen, 0*0346 

 for oxygen, and I *OO2 for carbon dioxide. 



I litre of sea water, salinity 3*5 per cent, will under the same con- 

 ditions contain 



1 1 'i c.c. nitrogen 



5 '8 ,, oxygen 



about 50 ,, carbon dioxide 



There is less nitrogen and oxygen because the absorption coeffi- 

 cients of the salt solution are lower than that of distilled water. The 

 enormous difference with regard to carbon dioxide depends upon the 

 fact that sea water is slightly alkaline, and takes up CO 2 to form car- 

 bonates and bicarbonates. Most natural fresh waters also contain 

 variable amounts of alkali, thus binding considerable quantities of 

 carbon dioxide with a very low tension of dissociation (Krogh [1904]). 



The total amount of dissolved gases can be extracted and deter- 

 mined by the mercury pump and subsequent gas analysis, but in order 

 to obtain all the carbon dioxide from natural waters in this way, it is 

 necessary to acidify (with dilute hydrochloric acid). The method is 

 cumbrous and difficult, and very great care is required to obtain accu- 

 rate results. It has been used in a few cases only for respiration 

 experiments. 



The dissolved oxygen can be determined accurately and easily by 

 titration after the method of Winkler (Cronheim, Bjerrum, Winter- 

 stein [1909]), the principle of which is to add manganese chloride and 

 caustic soda with potassium iodide, thus producing a sediment of man- 

 ganous hydroxide which almost immediately combines with the oxygen 

 present forming manganic hydroxide. When the liquid is afterwards 1 

 acidified with hydrochloric acid the hydroxides are dissolved and a 

 quantity of iodine set free, which is equivalent to the quantity of oxy- 

 gen formerly present and which is determined by titration with a solu- 

 tion of sodium thiosulphate. 2 



1 The interval can be of any desired length from two hours upwards (Bjerrum). 



2 Various methods are in use for standardizing the thiosulphate solution. The writer 

 has found the best plan to be to titrate distilled water saturate^ with pure air at 15 which 

 is known to contain 7*32 c.c. oxygen per litre (Fox [1907])^ 



