130 



Mr. H. Hartridge. Factors Affecting the [Nov. 1, 



question to require consideration was the cause of the divergence, but the 

 effect this might have on the accuracy of the carbon monoxide estimations had 

 also to be investigated. The differences that may exist between the readings 

 of two observers may be seen from fig. 1. 



On the left are plotted the observations and calibration curve obtained 

 from them for the observer G. W. On the right is drawn a portion of my 

 own calibration curve taken from the previous paper.* On comparing 

 corresponding points on the two curves it will be seen that the bands appear 



80 



40 



• 



• 



/ o 



Km / 



■ 



/ * 



, ••/•-M / 



* / / 

 • / »-M / 



0/ / 







-20 +20 60 100 



Fig. 1. 



•, initial observations ; M, mean of several readings ; O, final observations. 



everywhere displaced by about 20 scale divisions (10 A.U.). The initial 

 observations of G-. W. are given in the table below and are shown plotted as 

 black dots. 



But to ascertain how far subsequent readings could be relied on to give 

 the actual percentage saturation of a sample of blood, a further series of 

 isolated readings were made. These are shown plotted in fig. 1 as white 

 dots, and are given in the table side by side with the actual percentage 

 saturation present ; it will be seen that the latter readings have an average 

 error of 1*1 per cent. These figures show that although a considerable 



* ' Joum. Physiol.,' 1912, vol. 44, p. 23. 



