590 [June 11, 



Though thinking it not unlikely that this was the true explanation 

 of the influence of temperature on coagulation, I thought it worth 

 while to subject the matter to experiment. For that purpose I kept 

 the blood of a horse fluid by means of a freezing-mixture, and after- 

 wards by ice-cold water ; and when the corpuscles had subsided from 

 the upper part of the blood, I cautiously added to the liquor sanguinis 

 extremely dilute ice-cold acetic acid till it was of distinctly acid reac- 

 tion, the liquor sanguinis being of a colour that permitted the delicate 

 application of test-paper, which is impossible with red blood. By 

 this means any free ammonia which the fluid might have contained 

 must have been neutralized ; yet so long as it was kept in the cold it 

 continued fluid, but when brought into a warm room, coagulated 

 just as a specimen which had not been acidulated. Thus, when there 

 could be no free ammonia in the liquor sanguinis at all, it was still 

 affected as usual by temperature. 



This experiment may not be satisfactory to all minds, though I 

 confess it appears so to me ; and as this is a point of very great 

 interest, I have sought in another way for evidence regarding it. 

 First, however, I will mention an experiment which will not at once 

 appear to bear on the question of temperature. I drew out a fine 

 glass tube in such a way as to produce a fusiform receptacle continued 

 longitudinally each way into a tube of almost capillary fineness for 

 about two inches, which again expanded at the end, as represented in 

 fig. 3. Having squeezed out a drop of blood from my finger, I 



Fig. 3. 



sucked up a portion into the tube till the receptacle A and its capil- 

 lary extensions were filled. I then broke off the expanded ends, and 

 placed the little tube thus filled, B, in a bath of the strongest liquor 

 ammoniac. Here certainly the blood was in circumstances in which 

 it could not lose ammonia, but where, any change in its amount must 

 be by way of increase, and yet I found, on opening the receptacle by 



