PERFORMANCE OF THE EXPERIMENT 189 



\Ye have described the performance of the experi- 

 ment as it is carried out at the present time. 

 Previously it was usual to make use of the biuret 

 reaction for proofs of decomposition of albumen. To 

 10 c.c. of the dialysate 2*5 c.c. of a 33 per cent, 

 solution of caustic soda were added, and this was then 

 covered with a layer of very dilute copper sulphate 

 solution. (See here p. 155.) If a violet to red ring 

 appeared, the reaction was recorded as positive. 



The biuret test has been entirely given up for the 

 ninhydrin test, because the majority of observers have 

 a difficulty in distinguishing with certainty a feeble 

 biuret reaction. Those, however, who are capable 

 of distinguishing a biuret reaction, however slight, 

 should adhere to this test as well under all conditions. 



SOURCES OF ERROR ix THE DIALYSATIOX PROCESS. 



There are many possibilities leading to erroneous 

 results. It is best to consider them from the point 

 of view of utensils employed and manipulations 

 adopted, and to refer again to the sources of error 

 mentioned in the description of the method. 



(i) Tubes.--\Ye take it for granted that all tubes 

 are thoroughly and accurately tested before anything 

 else. On the average about 20 to 30 per cent, of the 

 dialysing tubes supplied by the firm Schleicher and 

 Schiill will be useless, because there are nearlv 



