The Hydrogen-ion Concentration, etc., of Sea-Water. 35 



Henderson and Cohn state that if the CO2 tension is maintained 

 constant (at about 0.0004 atmosphere) the Ph changes from 8.03 at 

 30.73 salinity to 8.1 at 34.6 S. We were unable to confirm this result 

 with the hydrogen electrode. Henderson and Cohn used the indicator 

 method of Sorensen and Palitzsch, and the difference between succes- 

 sive standard tubes as recorded in several of the papers of these authors 

 is Ph = 0.1. The tubes were exposed to the air, but were occasionally 

 renewed. Their results were sometimes expressed in figures of four 

 decimal places, but this probably was not intended to have more than 

 mathematical significance. In fact, we found that, with the most 

 accurately measured sealed tubes placed in a colorimeter that brings 

 the centers of the tubes together on a sharp line, we could not dis- 

 tinguish with certainty any difference smaller than about Pg 0.02, even 

 in the most sensitive region of the range of the most brilliant indicators. 

 The fact that Henderson and Cohn observed Ph 8.06 at S 31.58, 8.07 

 at 32.05, and 8.06 at 32.45, indicates that they made small errors in the 

 estimation of Ph or CO2 tension or temperature. Sorensen and Palitzsch 

 give the salt error of their indicators at 20 and at 35 salinity, and 

 Henderson and Cohn may possibly have attempted to estimate the 

 change in salt error from S 30.7 to 34.6, but they do not mention the fact. 



We have recahbrated our salt errors about 50 times, but the different 

 determinations do not coincide within an error of less than 0.025 Ph- 

 On the average it was found to be necessary to subtract 0.05 from the 

 observed indicator Pg (sulfophthalein series) when the salinity was 

 increased from 30 to 37.7 in order to obtain the same result as with 

 the hydrogen electrode. If this correction is applicable to Henderson 

 and Cohn's indicator for this range (probably phenolphthalein) the 

 increase in Ph with rise in salinity from 31.58 to 34.6 is reduced to 

 0.2 and is about the limit of the smallest possible difference that can 

 be determined with the eye, whereas the range of salinity is as great 

 as is ordinarily met with in the larger oceans. We feel safe, therefore, 

 in assuming that the influence of salinity on the relation of CO2 tension 

 to Ph at any specified temperature in negligible. In fact, Henderson 

 and Cohn do not state the variation in temperature. In our experi- 

 ments the temperature was observed within 0.05° and controlled within 

 0.2° during the majority of the experiments and to within 0.1° during 

 critical periods. 



In figure 6, the curves for 30°, 20°, and 10° are taken from the 

 means of many determinations, whereas the curve for 0° was extra- 

 polated. The results of Henderson and Cohn's experiments at 20° are 

 marked by crosses. Their results at higher CO2 tensions more nearly 

 fall on a smooth curve, but it is impossible to compare them with 

 hydrogen-electrode determinations, because the hydrogen electrode 

 is not very reliable in low partial pressures of hydrogen, even though a 

 correction be applied for change in hydrogen pressure or concentration. 



