370 BACTERIOLOGY. 
duction the neutralizing value of the fatal dose increases 
steadily until it becomes five to ten times its original 
amount. 
In our experiments the greatest value for L+ was 
126, the least 27. As at six hours L+ was only 72 
and at twenty-eight hours only 91, we doubt whether 
L+ ever reaches above 150.1 When we seek to analyze 
the above-described process we find certain facts which 
seem partly to explain it. Experiments have shown 
that filtered toxin, preserved for any length of time in 
conditions under which access of air occurs, gradually 
loses in both its toxicity and neutralizing power, and 
that it loses more rapidly in the former property than 
in the latter. Thus, while the fatal dose of a toxin pre- 
served for one year rose from 0.01 c.c. to 0.55 ce., it 
lost only half as much in neutralizing value, 1 unit 
neutralizing at first 1 ¢.c., at the end of the year 0.25 
c.c. These processes take place more rapidly at room- 
temperature than in the ice-chest, and in the incubator 
than in the room. 
In the fluid holding the living bacilli we have, there- 
fore, after the first few hours of toxin formation, a 
double process going on—one of deterioration in the 
toxin already accumulated, which tends to increase the 
neutralizing value of the fatal dose; the other of new 
toxin formation, which probably tends to diminish the 
neutralizing value. The chemical changes produced 
by the growth of the bacilli in the bouillon tend to aid 
one or the other of these processes, and so to make, from 
hour to hour, slight changes in the value of the fatal 
1 Li = fatal doses of toxin required to kill a guinea-pig in four days after 
having been mixed with one unit of antitoxin. 
Lo = fatal doses of toxin required to fully neutralize one unit of antitoxin- 
