554 THE POPULAR SCIENCE MONTHLY 



that the method was far from perfect. "What in the method caused so 

 great a variation? 



An examination of animals that had died from small amounts dem- 

 onstrated that death. had been due to coagulation and to a consequent 

 blocking of the blood stream. They then set about for a solution of the 

 more serious problem — the prevention of coagulation. First, tests with 

 various salts were made. But these were found to be of no service since 

 a salt in order to be a non-coagulant had to be of sufficient strength 

 itself to be toxic. 



A second study hit upon an ingenious method of procedure. It has 

 long been known that the blood of animals ingested by leeches is pre- 

 vented from coagulation (in the body of the leech) by an anticoagulant. 

 Haycroft demonstrated that an alcoholic extract of the buccal cavity of 

 leeches injected into the arteries of rabbits or dogs prevents coagulation 

 of the blood, and at the same time is productive of no observable in- 

 jury. Joffroy and Serveaux determined upon testing its powers to 

 prevent coagulation in a normal salt solution. 



It is evident that the addition of so complex a substance as leech- 

 extract to the blood of an animal must be made only with the most 

 careful control. Two things were demanded of it: (1) It must serve its 

 purpose, in this case prevent coagulation; (2) it must in no way injure 

 the animal. In testing for its injurious effects it was found that the 

 injection of enormous quantities 4 produced no injury. Since no coagu- 

 lation followed they were in possession of an anticoagulant by the aid of 

 which they could test the toxicity of a substance added directly to the 

 blood stream. 



Later Experiments and Units of Measure 

 (a) The Experimental Toxic Equivalent 



Joffroy and Serveaux established as a convenient unit of measure the 

 amount of alcohol that would kill per kilogram while the experiment 

 was in progress. This they called the experimental toxic equivalent. 

 This limit, as is evident, has the advantage of greater rapidity than that 

 (36 hours) used by Dujardin-Beaumetz and Audige. By injecting the 

 alcohol and this anticoagulant into the blood, in a series of eight ex- 

 periments, the following amounts of alcohol were found sufficient to kill 

 per kilogram of body weight: 12.65 c.c, 12.18, 11.69, 10.32, 10.51, 

 11.99, 12.48, 11.70. This series shows a striking regularity with ex- 

 tremes varying only between 10.32 and 12.65 c.c, variations which would 

 be readily accounted for by differences in age, race and the like of the 

 rabbits used. 



For ethyl alcohol, then, these authors have demonstrated that 11.69 



4 1,185 c.c. — nearly 600 grams per kilogram — was injected; of this some was 

 lost, but 425 grams per kilogram remained. 



