296 BACTERIAL VARIATION 



the original environment was restored. We must nevertheless guard against a too 

 literal interpretation of this difference, for as we have pointed out above, bacterial 

 generation under experimental conditions is so rapid as to compensate for a very 

 low mutation rate (the ratio of the number of mutations to the total number 

 of cell divisions). It follows that if mutations in the reverse direction were as 

 frequent as those in the direction of an observed variation, variation and rever- 

 sion might be accomplished with such speed as to suggest a temporary adaptation. 



Mutations are not equally rare in all bacterial strains. Indeed, some produce 

 variants so constantly and in such relatively large numbers that they are often 

 designated as " unstable " strains. Deskowitz (1937) concluded from a study 

 of unstable colony variants of Salm. typhi-rmirium that unstable strains differed 

 from stable only in having a mutation-rate of the order of 1 per cent, or more. 

 The existence of these " unstable " strains raises a practical point of some 

 importance in distinguishing adaptation from true mutation. 



We have seen in Chapter 3 that the typhoid bacillus, though normally dependent 

 for its growth on the presence of tryptophan, can be trained to synthesize this 

 essential metabolite from ammonium salts ; and in Chapter 6 there are numerous 

 examples of habituation of bacteria to growth in otherwise inhibitory concen- 

 trations of various antibacterial agents. These changes, in most cases permanent, 

 are excellent examples of impressed variation in biochemical constitution. A few 

 examples from recent work on the vitamin metabolism of bacteria and yeasts 

 will serve to show the effects of such training, and the degree of success which 

 attends it. 



Wood, Anderson and Werkman (1938) trained a strain of propionic acid bacteria to 

 dispense with thiamin, and later showed (Silverman and Werkman 1939) that the trained 

 cultures were synthesizing a substance that had the biological qualities of thiamin. 

 Leonian and Lilly (1942, 1943) induced several strains of the yeast Saccharomyces cerevisioe 

 to grow without one or more of certain essential vitamins, including thiamin, pyridoxin, 

 inositol, pantothenic acid, and in some cases biotin. At least two of the variants grew 

 without all five vitamins, and in some cases the growth in the absence of one vitamin 

 induced the power to dispense with other vitamins as weU. Some variants reverted 

 easily, but many of them reverted only after six months' subcultivation on a medium 

 containing all the vitamins. 



Koser and Wright (1943) trained four strains of dysentery baciUi to dispense with 

 nicotinamide. The variants could be obtained either by serial subculture in a glucose 

 amino-acid medium containing diminishing quantities of nicotinamide ; or by incubation 

 of a large inoculum in a nicotinamide -free medium. Under the latter conditions the 

 proportion of variants to the total of viable cells increased during incubation. The 

 variants resembled the parent strains both antigenicaUy and in their gross fermentation 

 reactions, but never developed as luxuriantly as in media containing optimal amounts 

 of the vitamin. Nevertheless, they had acquired the power to synthesize a substance 

 physiologically equivalent to nicotinamide, for filtrates of their cxiltures in vitamin-free 

 media supported the growth of strains of bacteria known to require nicotinamide. 



Whereas the typhoid bacillus appears to acquire the ability to synthesize 

 tryptophan with relative ease, its normal inability to attack lactose seems to 

 depend on a more fundamental peculiarity of cell organization, since it is extremely 

 difficult to produce lactose-fermenting strains of this organism, even by prolonged 

 " training " in media containing this substrate as the main available source of 

 carbon. Twort (1907) has succeeded in producing one such strain ; and we may 

 perhaps regard the extreme rarity with which this change has been induced as 



