304 BACTERIAL VARIATION 



The S — >■ R variation occurs frequently under ordinary laboratory conditions of 

 cultivation, and may be readily induced by tlie methods that we shall shortly 

 describe. The reverse change (R — > S) seldom, if ever, occurs under the ordinary 

 conditions of cultivation and it is very difficult, though not impossible, to induce 

 it by any specific stimulus, when the original S — > R change has been complete. 

 The evidence indicates that this S — > R variation, at least as a quality of a culture 

 as distinct from that of a component cell, is not a sudden " all-or-none " process, 

 but a gradual or step-like change, so that intermediate SR forms appear between 

 the typical S and the fully degraded R. In these partially degraded SR variants 

 reversion to the normal S form may be more easily induced. 



Prolonged growth of a normal smooth strain in any of the ordinary fluid media 

 of the laboratory, followed by plating on ordinary agar, will usually result in the 

 appearance of a proportion of rough, or partially rough, colonies. A bacteriophage 

 that causes lysis of the normal smooth strain provides another, and very potent, 

 method by which this change can be induced (see Chapter 11). The contamination 

 of a bacterial culture with a bacteriophage is not, however, an entirely desirable 

 procedure for this particular purpose ; and the best method available is that 

 introduced by Griffith (1923), who showed in the particular case of the pneumococcus, 

 that rough variants could readily be produced by growing the normal smooth form 

 in the presence of an antiserum acting on the type-specific capsular polysaccharide. 

 This method, the efficacy of which has been repeatedly confirmed by other workers, 

 appears to be of quite general applicability. 



The induction appears to act partly by selection. In a broth culture baciUi with a 

 great deal of the antigen in question will be flocculated in a deposit by the antibody, 

 and variants with none or less of it will tend to remain in suspension, so that a sample 

 from the upper part of the fluid will contain a relatively high proportion of the variants. 

 A method that can be used with motile flagellated bacteria is growth in a semi-soUd agar 

 medium containing flagellar antibodies ; those bacteria that by reason of their motility 

 spread from the original inoculum through the semi-sohd agar will tend to be those with 

 little of the homologous flagellar antigen. By this second method Gnosspehus (1939) 

 induced flagellated variants in diphasic salmonella bacilli possessing antigens different 

 from the original characteristic " type " or " group " antigens. The variation was irre- 

 versible, and though the new antigenic types retained their capacity for diphasic variation, 

 it was found that only the phase which had undergone impressed variation had altered. 

 Thus, the " new " group phase alternated with the " original " type phase, and a " new " 

 type phase alternated with the " original " group phase. Eriksson and Malmstrom 

 (1939) induced a simUar variation in Salm. newport, and found that the " new " flagellar 

 variant had acquired an antigen of the /S type; i.e., an a — > P variation (see p. 716) 

 had been induced. These organisms were originally diphasic. Starting with the mono- 

 phasic Salm. paratyphi A, Bruner and Edwards (1941) induced four variants, one in the 

 original specific phase, one corresponding to the group phase of diphasic salmoneUae, 

 and two new phases with antigenic components unlike any. described for the salmoneUae. 



It is clear that induced variations, which in this case revealed not only a potentiality 

 for diphasic variation, but also two hitherto undiscovered antigens, off"ers a means of 

 exploring hidden antigenic relationships within groups of bacteria. 



As an example of a phage-induced variation, we may cite the production from typhoid 

 baciUi with Vi, O and R antigens, of variants which have either Vi and no O antigen, 

 or O antigen without Vi (see, e.g., Kauffmann 1936, Craigie and Brandon 1936). 



We are then, in the change from S to R, dealing with a striking example of 

 a variation, of a very definite type, that can be induced at will by certain specific 



