CH. v] VARIATIONS IN FERMENTING POWER 63 



their contact with the new sugar, such changes being pre- 

 paratory to the acquisition on their part of new fermenting 

 powers. The observation already referred to, that a very 

 rapid increase in the number of fermenting organisms occurs 

 simultaneously with the appearance of acidity, tends to 

 support this hypothesis. Penfold has, however, disproved it 

 by experiment. He observed that B. typhosus when grown 

 in dulcite broth gradually acquired the power to ferment the 

 sugar. After several days had passed, plating out on dulcite 

 agar showed that 95 per cent, of the strain were dulcite fer- 

 menters. He found that subcultures from the non-fermenting 

 colonies into dulcite broth took the same length of time as the 

 original stock to produce fermentation, thus showing that 

 they had not undergone any preparatory change during the 

 first incubation period. 



In the second place, if the development of the new fer- 

 menting power is dependent in the first instance upon the 

 occurrence of a spontaneous "fluctuating" variation in the 

 required direction, and such variations are infrequent, an 

 interval of uncertain length must necessarily intervene be- 

 tween the commencement of the experiment and the appear- 

 ance of the variant which is to give rise to the new strain. 

 The fact that the length of the incubation period, whenever 

 a certain organism is " trained " to ferment a certain sugar, is 

 fairly constant, disposes of this argument. 



In the third place, the original non-fermenting strain might 

 contain a very few fermenting individuals, in insufficient 

 numbers, however, to give any evidence of their presence. 

 These few "fermenters" would possess an advantage over the 

 "non-fermenters" and multiplying more rapidly would, in time, 

 outnumber the latter, but a certain period would necessarily 

 elapse before they gained the ascendancy. Inasmuch, however, 

 as the original strain can be made in the same way to ferment 

 a number of different sugars, the strain must, on this hypo- 

 thesis, contain at one and the same time fermenters of each of 

 these different sugars. Even if this be granted the hypothesis 

 offers no explanation of the fact (illustrated by the behaviour 

 of B. typhosus in dulcite and in lactose respectively) that in the 



