135 



of (haemoglobinophilic) bacteria it is important to know whether 

 we can count on the grouping of the strains according to their 

 behaviour in the symbiosis test, which we get for example 

 by employing the air coccus as growth-promoting organism, 

 being the same as when the anthrax bacillus, for example, is 

 used. If it is always the same growth-promoting substance 

 we have to deal with, this must necessarily be the case. But 

 if the substances which different species of bacteria form 1 are 

 distinct, this will be shown by the fact that one will act 

 particularly upon some strains of haemoglobinophilic bacteria, 

 and another particularly upon other strains. Such a condition 

 of affairs would indeed make the symbiosis test more difficult 

 to apply. 



To investigate this, I chose, with the help of the big 

 symbiosis experiment described above, 9 of the bacterial cul- 

 tures employed which differed considerably from one another 

 both in their growth-promoting action on the same strain of 

 Pfeiffer's bacillus, and in their general biological characters. 

 The idea was this: if only one growth-promoting substance 

 for Pfeiffer's bacilli is produced by bacteria, then the variable 

 intensity of the increase of growth of cultures of one strain 

 of Pfeiffer's bacillus must depend upon the fact that the sub- 

 stance is produced in unequal amounts, and if we test the 

 same 9 strains against any other Pfeiffer's bacilli (by which 

 is here understood merely a haemoglobinophilic bacterium 

 which can give the symbiosis reaction), these strains must, 

 as regards growth-promoting effect, form the same series among 

 themselves as before. If the effect of a strain „A" on a strain 

 of Pfeiffer's bacillus is greater than that of a strain ,,B", then 

 the effect of „A" on any strain of Pfeiffer-bacillus whatever 

 must be greater than that of „B". But if the different strains 

 form 1 different growth-promoting substances of which a par- 

 ticular one acts on some strains of Pfeiffer's bacillus, and ano- 

 ther especially on others, then the activity series will be dif- 

 ferent according as they are tested against one or other strain of 

 Pfeiffer's bacillus, provided strains are employed which are as 

 different as possible among themselves. 



As growth-promoting organisms 5 strains were used, which 

 had shown this property to a marked degree in the previous ex- 

 periment (air coccus, Staphylococcus aureus, haemolytic Streptococ- 



