Ralph W. Webstee 



were made of defibrinated blood and of isosmotic salt solutions in various dilutions. 

 In some cases equal parts of each (blood and salt solution isosmotic with blood) were 

 taken. In other cases, 2, 3, 5, and 10 parts of solution to 1 of blood were used. 

 These mixtures were placed in dishes containing, approximately, 25 c.c, and allowed 

 to stand for intervals of 1, 3, 6, and 24 hours. At the end of each interval readings 

 were taken by the haematokrit method, using one revolution. per second for 3 minutes. 

 By this method the percentage relation of corpuscles to serum was obtained. These 

 experiments can have no further value than a comparative one, inasmuch as the osmotic 

 pressure of the mixtures was not determined, and thus the concentration of the solution 

 acting on the corpuscle was unknown. However, as solutions of salts isosmotic with 

 the blood serum and, therefore, with one another were used, we may readily show 

 whether osmotic pressure differences account for phenomena noted or whether we are 

 dealing with specific ionic effects. Below will be found a table embodying the results 

 of these experiments: 



TABLE I 



A study of this table shows two facts quite clearly. The first is, that we have 

 two classes of substances represented here. The one class, to which the corpuscular 

 membrane is permeable, embraces NH4CI and urea; the other class, to which the 



109 



