173 



its agglutination reactions. Further, in Jan. 1920 it formed colonies 

 with a finely granular surface, while in September they were smooth. 

 A comprehensive mutation must therefore have arisen, or else a 

 contamination of the culture or its substitution by another must 

 have taken place. 



These four cases are the only ones in the material included 

 in the table for which it is necessary to have recourse to the three 

 possible explanations just given. This must be considered very 

 satisfactory because it must not be forgotten that the sole guarantee 

 of purity of the majority of the cultures consists partly in the fact that 

 each strain was cultivated from an isolated colony on the original 

 plate, and partly that there is good reason to suppose that contamina- 

 tions with other bacteria than Pfeiffer's bacillus would practically 

 always be detected and got rid of by plating out. 



An absolutely pure cultivation e. g. by Burri's or 0rskov's 

 (1,2) method, would undoubtedly be fraught with considerable dif- 

 ficulties on account of the smallness and slight resistance of the 

 bacillus, particularly if it had to be undertaken on an extensive 

 scale; and I did not think that the relative certainty which could 

 be attained by the usual plating out of each strain a number of 

 times would repay the labour entailed and the considerable expen- 

 diture of material. It is impossible to deny that during the ex- 

 tremely numerous inoculations a mistake may have been made in 

 one or two cases. In such an extensive statistical investigation as 

 the present a few irregularities are only to be expected. 



In series VI the positive reactions were on the whole rather 

 weak. Several of them could only just be recognised as positive. 

 I was unable, at the time, to subject these cultures (which as 

 will be remembered, were subcultured 98 times in rapid succession) 

 to a fresh examination on other media. It is however highly pro- 

 bable that the weakly developed indol formation or its complete 

 absence in some cases, must be accounted for by the constitution of 

 the peptone. 



The significance of the kind of substance we give Pfeif- 

 fer's bacillus to manufacture indol from, is well seen in the 

 series IX— XII all carried out at the same time. A particular 

 „peptone u was used in each of these series and care was taken 

 to choose peptones as different from one another as possible. 



Witte peptone, as far as is known, is a peptic product 

 and as such it principally contains albumoses. Peptone Gha- 

 potaut is considerably further broken down and contains about 

 20<y amino-acids (see Stickel & Meyer). Trypsin-digested 

 casein (here referred to as „casein-tryptone") is also extensi- 

 vely split up and contains in particular, an unusually large 

 proportion of free tryptophane (see Cole & Onslow). Lastly, 

 erepton is said to be completely broken down into animo-acids. 



