W. W. Palmer 125 
and protected from light. Attention should be called to the fact 
that the scales of the Hellige colorimeter are often inaccurately 
placed. The standard solution must be checked against itself 
and the scale adjusted so as to read 100 when the color in the 
cup and wedge match. We have found more difficulty in ob- 
taining accurate checks with. this instrument than with the 
Dubosceq or Kober colorimeters; but with care the error should not 
exceed 2 per cent. 2 cc. of solution is sufficient for the Hellige 
cup, hence the Sahli pipette, which contains 0.020 ce. of blood, 
may be used with this colorimeter. As large errors have been 
found in the calibration of the Sahli pipettes, it is pty to 
recalibrate them before using. 
Accuracy of the Method.—The accuracy of the method de- 
pends to a large extent on the care of the operator in carrying out 
the various details of the technique. The several steps involved 
in the method are common chemical procedures with known 
limits of accuracy. With care duplicate determinations are close 
and the error should not be more than 1 per cent. A series of 
ten determinations was made on the same blood. In eight of the 
ten, the first reading of the colorimeter was in every case. exactly 
the same. In the two remaining a difference of 0:1 of a division 
on the colorimeter scale occurred. | 
In the following table are presented a few eecenere by 
the method described using the Duboscq and Hellige colorimeters 
and comparing with the values estimated from the oxygen 
Hemoglobin determinations. ggeceiag 
eels | Omyesn | Dubosea and | Gifterence. 
Duboseq. Hellige. capacity. capacity. 
per cent per cent per cent 
1 105.5 “104.8 106.5 E10 0.94 
2 95.2 92.0 94.2 =e) 1.06 
3 6128;3 125.0 129.5 +1.2 0.93 
4 *112:5 115.2 tie? —0.8 0.68 
5 116.2 116.3 116.4 Ht).2 0.17 
6 79.4 84.0 80.5 ed 1 1.36 
7 119.2 115.0 119.4 Pere 0.17 
8 84.0 85.0 apes) UES 0.95 
9 82.0 82.0 82.6 +0.6 0.73 
10 108.6 a 108.1 025 0.46 
