Studies in Bacterial Variability. 



55 



re-examined the question in another relation (1918) (3) with four strains of 

 B. typhosus. A rabbit was immunised successively with three of the strains, 

 and on each occasion the titre of its serum rose to a higher level for the strain 

 used as antigen on that occasion, than for either of the other three strains. 

 Furthermore the four strains were seen to belong, two and two, to two sero- 

 logical groups which were widely differentiated from each other by their 

 relative sensitiveness to agglutination by different serums.* 



Somewhat similar serological differences are now well known in certain 

 other organisations, e.g., Meningococcus, Pneumococcus ; and they may at times 

 be very marked indeed, so that particular strains are described as " inagglutin- 

 able " — for example inagglutinable strains of B. typhosus. Yet it is found 

 that on sub-culture these inagglutinable strains will sooner or later yield 

 cultures which give good agglutination. Thus the inagglutinability appears 

 to be only a phase, or temporary character. The inagglutinable form of the 

 organism cannot, therefore, on the existing evidence he regarded as mutant, in the 

 proper sense of the term ; but it presents an example of bacterial fluctuation. 



Since the observations just referred to (1918) (3) raised a number of 

 interesting questions the subject has been re-examined in detail during the 

 progress of the present investigation by A. D. Gardner and Ainley Walker 

 (1921) (4) who obtained inagglutinable strains of B. typhosus, and compared 

 them serologically with ordinary " good agglutinators." The existence of the 

 two serologically different types of B. typhosus was fully confirmed. It was 

 shown that they corresponded to a motile and a non-motile phase respectively 

 of the bacillus. And it was further shown that from a single culture of the 

 bacillus colonies could on occasion be isolated by plating which would on 

 cultivation give rise to populations differing as widely, both serologically and 

 as regards motility, as any strains obtained from different sources. 



The serological difference between the agglutinable (motile) form and the 

 inagglutinable (non-motile) form of the bacillus was an antigenic difference. 

 For if a serum were prepared with the inagglutinable form, it agglutinated its 

 own bacillus and other inagglutinable strains quite well, and to high titre. 

 But in contrast with the rapidly-formed, large and fluffy flocculi produced in 

 an ordinarily well-agglutinating culture, the clumps produced in inagglutinable 

 suspensions are always small, compact and slowly formed. 



On this account, and in view of what is to follow, I propose the term 

 dys-agglutinablcf as a more appropriate designation than " inagglutinable " for 



* A more complete record of this experiment was published by Gardner and Ainley 

 Walker (1921) (4). 



t This is a hybrid word ; but so are a good number of other accepted and useful 

 terms. 



