184 INFECTION AND IMMUNITY. 



protected the animal against the ten fatal doses of 

 toxin contained 1/10 of an immunity unit, and 

 from this result the number of units in a cubic 

 centimeter could be calculated. This method in- 

 volved the use of toxin as the standard by which 

 the value of the antitoxin was measured, and it 

 was found to be unreliable. A toxin degenerates 

 rather rapidly, retaining at the same time its 

 binding power for the antitoxin; hence two tests 

 made with the same serum two months apart 

 might indicate different antitoxic values for the 

 serum. Also 10 fatal doses of one toxin often re- 

 quired more antitoxin for neutralization than the 

 same quantity of a second toxin. These phenomena 

 are due to the formation of toxids. (See next 

 chapter.) 



standard On account of these sources of error, Ehrlich 

 ns * devised a new method in which a standard anti- 

 toxin or test-serum is used as the starting point 

 for the valuation of a new serum. The test-serum 

 used at the Koyal Prussian Institute for Experi- 

 mental Therapy at Frankfurt, of which Ehrlich 

 is the director, is a dried and powdered serum of 

 such strength that 1 gram contains 1,700 immun- 

 ity units; i. e., 1/1700 gm. would protect a guinea- 

 pig against 100 fatal doses of a diphtheria toxin. 5 



5. In Germany the various serums are prepared by pri- 

 vate individuals or corporations and manufacturers are re- 

 quired to send a sample of every lot of serum intended for 

 the trade to the Frankfurt Institute that its exact value may 

 be determined. Each bottle eventually receives a stamp sig- 

 nifying the value in antitoxin units of the contained serum. 

 Moreover, samples of every lot of serum are retained in the 

 institute, and from time to time these are tested ; and when 

 it is found that the samples have degenerated beyond a cer- 

 tain value the order is sent out to call in all serum belong- 

 ing to the degenerated lot. When a manufacturer thinks 



