THE SMOOTH -^ ROUGH VARIATION 301 



it was also high in five of eight intermediate strains, and in one of seven rough strains. 

 Among the coli-typhoid group, indeed, the production of large amounts of polysaccharide 

 mucoid substance is sometimes associated with no change, or with a fall, in virulence. 

 Mucoid variation is often favoured by growth at 25° C. instead of 37° C. (Birch-Hirschfeld 

 1935, Morgan and Beckwith 1939, vor dem Esche 19406). Vor dem Esche (1940a) noted 

 the appearance of mucoid variants in the stool of a woman who had been treated by 

 inoculations with an autogenous vaccine during convalescence from paratyphoid infection. 



In many of these instances the change from smoothness to roughness is also 

 associated with a complete or partial loss of virulence, a factor which is clearly 

 of the first importance from the point of view of the medical bacteriologist. 



Again (see Chapter 11) the S — > R variation is usually associated with a change 

 in sensitivity to the lytic action of various strains of bacteriophages, the filtrable 

 viruses that propagate on, and at the expense of the bacterial cells. 



Changes that are exactly similar have been described in Pasteurella (de Kruif 

 1921, Webster 1925), many other bacilli of the coli-typhoid group (White 1926), 

 Staph, aureus (Bigger et al. 1927), and in a large number of other species (see, for 

 instance, Hadley 1927). We may, indeed, regard it as a variation to which most 

 if not all bacteria are subject. 



The S — > R variation affords an excellent example of a correlated variation. 

 If we accepted the change in colony form as the essential criterion of the change 

 from roughness to smoothness, we should note that the rough variant differed 

 from the smooth form in the following characters : 



(1) Roughness or granularity of colonies. 



(2) Instability in saline. 



(3) Loss of the antigenic component characterizing the surface of the bacterial 

 cell in the normal smooth form, whether this component is normally present in 

 the form of a definite capsule or not. 



(4) Loss of virulence, partial or complete. 



(5) Altered sensitivity to various bacteriophages. 



In accepting such a list of correlated characters we are, however, met with the 

 difficulty that the variations observed in certain bacterial species fail to fall into line. 



For instance, the normal, virulent, capsulated form of the anthrax bacillus 

 gives a flat, uneven, filamentous colony, while the avirulent, non-capsulated variant 

 is raised, circular and smooth (Preisz 1904, 1911, Eisenberg 1912). The ha^molytic 

 streptococci appear also to form an exception to the general rule. The colony 

 given by the normal virulent form is finely granular, not smooth, and the avirulent 

 variant, which has lost its type-specific antigen, is smoother, not rougher than the 

 normal form. For this reason Todd (1928), who first gave a detailed description 

 of these variants, used the term "matt" to describe the normal colony form, 

 and " glossy " to describe that of the avirulent variant. 



According to Dawson, Hobby and Olmstead (1938) the range of variation possible 

 between highly virulent and non- virulent forms of streptococci is much wider than that 

 indicated by Todd's matt — >- glossy. They described in a number of hsemolytic strepto- 

 cocci four types of variants— a capsulated M variant, forming large, watery, mucoid 

 colonies, an MS variant, forming colonies corresponding to Todd's " matt " variants, a 

 non-capsulated S variant, corresponding to Todd's " glossy," and an R variant. With 

 progressive loss of characters associated with virulence, the variation proceeds as foUows : — 

 M — >. MS — >- S — >- SR — >- RS — >- R, where SR and RS represent variants intermediate 

 between S and R forms (see also Morison 1940, Seastone 1943). A similar series of variants 



