158 



Scientific Proceedings (54). 



there is a marked increase in the quantity of hemoglobin per 

 corpuscle. That this is an actual increase and not a simulated 

 effect due to the presence of some substance with a greater light 

 absorption than oxyhemoglobin is shown by the constant value for 



1 7l 



Iog 7? 



in normal blood and in the blood of pernicious anemia. 



Closely related to the color index is the actual color of 

 the red blood corpuscles. It is well known that the color of 

 erythrocytes in single layer viewed microscopically in transmitted 

 light is yellowish green. The blood itself is deep red in color, 

 and the reddish tinge becomes noticeable microscopically when 

 several superposed corpuscles are viewed in transmitted light. 

 The explanation of this phenomenon becomes apparent, I think, 

 when one studies the spectrum of a single layer of corpuscles as 

 compared with several superposed layers. The absorption curve 

 of oxyhemoglobin presents a minimum in yellowish green (560 nn) 

 and a region of least absorption in red (650^-660^). In the 

 oxyhemoglobin spectrum k er is much greater than k r , con- 

 sequently on increasing the thickness of the absorbing layer the 

 intensity of the transmitted light diminishes much more rapidly 

 in yellowish green than in red. If we regard only these two regions 

 in the spectrum we would have for the intensity of the transmitted 

 light in yellowish green /„/ = I gT e~ kgrl and in red I,' = I r e~ krl . 

 For I gr e~ kgrl >I T e~ krl a color change would be expected. This is 

 known as the principle of dichromatism. Whether a color change 

 occurs or not with the same light source would depend on the values 

 of k gr and k r and on /. The necessary conditions I gr > I r , k gr > k r 

 for oxyhemoglobin and / sufficiently small are all realized in the 

 case of red blood corpuscles viewed in daylight. This formulation 

 is only a very rough approximation. The exact formulation would 

 require integration of the intensities over the whole spectrum, and 

 this cannot be done at present as long as it is not known what func- 

 tions / and k are of X (the wave-length). However, several ob- 

 servations furnish strong support for the view that the color 



