23 
j _ Ehrlich ^ now made numerous researches in order to determine the 
stability of the antitoxin. ,, He had particular success with a serum 
' I containing a high percentage of glycerin. This gl}xerinated serum 
j showed exactly the same power of neutralizing toxine after one 3 "ear 
that it did at first. 
For a while this glycerin solution of antitoxic serum was used as the 
basis for standardizing serums. But Ehrlich showed that even giyc- 
erinated serum ma}" weaken and also the fact that because the neu- 
tralization point does not vary is no guarant}^ of the stability of the 
I test solution. It was therefore necessary to abandon the gljxeriimted 
serum as a method for preserving the standard unit. The principal 
I factors which cause the weakening of antitoxin are moisture, oxygen, 
I light, and heat. 
It is very easy to guard against the deleterious influences of light 
' and heat. It was therefore necessary onl}^ to pay particular attention 
to the first two agents. 
The serum was therefore reduced to a powder under proper precau- 
tions and this dried blood serum preserved in a small glass apparatus 
consisting of two parts connected b\" means of a glass tube. The 
serum is placed in one portion of the apparatus and phosphoric anh}^- 
drid, which is a veiy powerful dely^drating substance, in the other. 
The air is exhausted as much as possible by means of a high vacuum 
and the apparatus hermeticall}" sealed. After a few days the acid will 
have taken up all the moisture from the serum, and the tube connect- 
ing the two parts may then be melted and sealed off, separating the 
acid from the serum. The dried serum is now contained in a vacuum 
tube and preserved in a cool place in absolute darkness. 
In this way Ehrlich prepared a great number of these little tubes, 
each one containing a dried serum whose value was accurately deter- 
mined. Each tube contained 2 grams of the dried serum, repre- 
senting 1,700 immunity units in each gram. Every two or three 
months one of these tubes is carefull}- opened and the contents dis- 
solved in 200 c. c. of a 10 per cent solution of sodium chlorid and glyc- 
erin mixture, the mixture containing from 50 to 80 per cent of 
glycerin. One c. c. of this gl}"cerinated serum represents then exactly 
seventeen times the normal strength; and therefore if we dilute 1 c. c. 
of this gl}"cerinated serum with 16 parts of water, each cubic centi- 
meter of the dilution contains one immunity unit. 
Although the serum is used as a basis for standardizing the strengths 
of other sera, the toxine still pla 3 ^s a very important role in the process 
of testing, as it is manifestly impossible to estimate the strength of an 
« Ehrlich : Die Wertbemessung des Diphtherieheilserums und deren theoretische 
Grundlagen. Klin. Jahrb., Jena, v. 6 (2), 1897, pp. 299-326. 
