326 A. C. R. Dean and Sir Cyril Hinshelwood 



have become able to utilize them after an initial reluctance 

 to do so. The difficult reversibility of such adaptations has 

 often been used as an argument to support the view that the 

 new forms are genetically different, containing presumably 

 DNA patterns modified by accidental mutation. But the 

 variable speed of reversion of cell adjustments in general 

 seems to invalidate at least this particular form of argument. 

 A mutant may indeed be more resistant than the original 

 form, but more proof than slowness to lose the resistance is 

 needed to establish that it is due to mutation rather than to an 

 adjustment induced by the drug. 



Our views on this question were summarized in a com- 

 munication to a previous symposium of the Ciba Foundation 

 (Dean and Hinshelwood, 1957a), and will not be repeated 

 here. They in no way under-rate the significance of informa- 

 tion stored in DNA patterns, but they allow considerable 

 scope for quantitative modification of cell characters by 

 enzymic adjustments, some of which may be established 

 quickly in new conditions, and reversed slowly on return to 

 the old conditions. 



Two developments will, however, be referred to here in 

 extension of what was said previously. Evidence cited in 

 favour of the view that there is a formation by direct adapta- 

 tion of D-arabinose-positive and lactose-positive forms of 

 coliform bacteria from the original "negative" strains was 

 called in question at the symposium itself. The basis of the 

 objections was the claim that during the sojourn of the 

 bacteria in the medium where the adjustment is taking place 

 platings of samples on to solid media containing D-arabinose 

 on lactose gave two types of colony clearly recognizable as 

 positive and negative. The answer to this objection, in our 

 view, is as follows. The adaptation times of individual cells 

 vary according to an approximately normal distribution. 

 Qualitative visual observation of a set of colonies with 

 diameters distributed in this way can be misleading, and 

 easily give the impression of "large" and "small" types, 

 whereas quantitative measurement shows a continuous dis- 



