158 Immunity 



index to its antitoxin-combining power, which, of course, must be 

 the foundation of the test. The toxin, under natural conditions, is 

 changed with varying rapidity into toxoids, of which he demonstrates 

 three groups prototoxoids, syntoxoids, and epitoxoids. The epi- 

 toxoids have a greater antitoxin-combining power than the toxin 

 itself, yet have no toxic action upon the guinea-pigs, hence cause 

 confusion in the results. 



To secure a satisfactory measure of the antitoxic strength of a 

 serum, it is therefore more important to first determine the antitoxin- 

 combining power of the toxin or toxic bouillon to be used, than to 

 determine its guinea-pig fatality, and this is what Ehrlich endeavors 

 to do. 



(B) Ehrlich's Method. In this method the unit is the same as in 

 Behring's method, but its determination is arrived at by a very im- 

 portant modification of the method, by which the standard of measure- 

 ment is a special antitoxin of known strength, by which the antitoxin- 

 combining power of the test toxic bouillon is first determined. Ehrlich 

 began by determining the antitoxic value of a serum as accurately as 

 possible by the old method and then used that serum as the standard 

 for all further determinations. The serum was dried in a vacuum, and 

 two grams of the dry powder were placed in each of a large number 

 of small vacuum tubes, connecting with a small bulb of phosphoric 

 anhydride. In this way the standard powder was protected from 

 oxygen, water, and other injurious agents by which variations in its 

 strength could be initiated. Periodically one of these tubes was 

 opened and the contained powder dissolved in 200 c.c. of a mixture 

 of 10 per cent, aqueous solution of sodium chloride and glycerin. 

 The subsequent calculations are all based upon the strength of the 

 antitoxin powder. In Ehrlich's first test serum 1 gram of the dry 

 powder represented 1700 units. Of the solution mentioned, 1 c.c. 

 represented 17 units; ^ c.c., one unit. 



Having by dilution 1 c.c. of the first dilution in 17 of water 

 secured the standard unit of antitoxin in a convenient bulk for the 

 subsequent manipulations, it is mixed with varying quantities of the 

 toxic bouillon to be used for testing the new serums, until the least 

 quantity is determined that will cause the death of a 250-gram guinea- 

 pig in exactly four days, when carefully injected beneath the skin 

 of the animal's abdomen. This quantity of toxin is the test dose. 

 If the toxic bouillon was "normal" in constitution, it should represent 

 100 of the least certainly fatal doses that formed the basis of the old 

 method of testing, but as toxic bouillons contain varying quantities 

 of toxoids it may equal anywhere from fifty to one hundred and fifty 

 times that dose. 



The test dose of toxic bouillon, having been determined, remains 

 invariable throughout the test as before, the serum to be tested for 

 comparison with the standard being modified. The calculation is, 

 however, different because the guinea-pig is receiving, not ten times, 

 but more nearly one hundred times the least fatal dose, and the quan- 

 tity of the antitoxic serum that preserves life beyond the fourth day 

 is itself the unit. 



Example: The sample of serum issued as the standard contains 17 

 units per cubic centimeter. Serum 1 c.c. + water 16 c.c. = 1 c.c. 

 is the unit. 1 c.c. of the dilution containing one antitoxic unit is 

 mixed with 0.01, 0.025, 0.05, 0.075, 0.1 c.c. of the toxic bouillon. All 

 the animals receiving less than 0.1 c.c. live. A new series is started, 

 and the guinea-pigs all weighing exactly 250 grams receive 1 unit of 

 the antitoxin plus toxic bouillon 0.08, 0.09, 0.095, 0.097, 0.1, 0.11, 

 0.12, etc. It is found that all receiving more than 0.097 die in four 



