20) os Electrometric Titration of Blood 
NaHCO; is changed to NaCl before the end-point is permanently 
ttained: i.e, for pH = 7 the ratio eo muse hoe 
attained; 7.e., for pH = e ratio aHGO, Must be about +. 
If, by washing out the COz gas, the H.CO; is reduced to’an in- 
finitesimal value, the NaHCO; is also reduced to approximately 
zero. | 
The effect of the titration therefore is to add sufficient HCl 
to change all the NaHCO; into NaCl. The phosphates of the 
serum have a negligible effect on the titration. They are nor- 
mally only about 34 themolecular concentration of the bicarbonate, 
and change of pH from 7.4 to 7.0 changes only about one-fifth 
of the phosphate present from NaHPO, to NaHe2PO,. The 
neutralizing power exerted by the serum proteins for the same 
pH change is calculated by Henderson to be equal to that of 
about 0.001 Nn alkali, while the bicarbonate is normally about — 
30 times as concentrated. Since the effect of the chief known 
buffers of the serum (aside from bicarbonate), v2z., phosphate 
and proteins, on the titration is calculated at only 749 and #5 re- 
spectively of the bicarbonate effect, the assumption appears 
justified that our titration measures the serum bicarbonate with 
a very slight error. As a matter of fact, we find that the amounts 
of acid required in our electrometric titration of normal human 
serum are equivalent to an alkali concentration of 0.030 N, 
which corresponds almost exactly ‘to the average bicarbonate 
concentration of normal plasma (65 volume per cent CO, = 0.029 Nn) 
found by Van Slyke and Cullen as a result of direct gasometric 
determination of the CO, portion of the bicarbonate molecule. 
In titrating plasma, I have used the rotating electrode (Me- 
Clendon, 1917, b) thus avoiding frothing and eliminating the 
necessity of pumping, but it was necessary to disconnect the 
electrode every time more acid was added. It seemed necessary, 
therefore, to devise an electrode into which acid could be run and 
through which hydrogen (or He-COz mixture) could be run with- 
out disconnection from the potentiometer. 
After some preliminary forms of electrode had been tried, the 
rotating electrode shown in Fig. 1 was found satisfactory. The 
electrode vessel is cylindrical (30 X 36mm.) with an opening 
10 mm. in diameter in one end and the other end cemented to a | 
cork pulley with sealing-wax, and the whole mounted in a wooden 
frame so as to rotate on a horizontal axis. Through the open end 
