DELAYED EXPRESSION OF MUTATIONS 465 



the inoculum. It is therefore convenient, when plotting a curve using data 

 from a number of experiments, to express the increase in terms of resistant 

 clones per 10 8 bacteria in the inoculum. This has been done in figure 1, where 

 (R 2 — Ri)/NiXl0 8 is plotted against growth (N 2 /Ni), time 1 being the time of 

 plating. Each point shown in figure 1 was obtained by averaging at least eight 

 independent estimates of increase in number of resistant clones and of growth. 



Two experimental procedures for estimating growth were used in this case. 

 One was that mentioned earlier, in which bacteria are washed from the plate 

 and the number determined by dilution and assay. The other was by direct 

 count of the numbers of bacteria in the developing microcolonies. 



For the latter purpose small numbers of bacteria were plated on agar, and 

 the plates were incubated at the same time and for the same period as those 

 that were to be sprayed with phage. These growth assay plates were then 

 chilled to stop division and examined under a high-power dry objective. The 

 numbers of bacteria in 50 microcolonies were counted and averaged. The 

 method gave accurate results up to an increase in number of bacteria of ap- 

 proximately 64 times. 



In figure 1, points obtained by calculating growth by means of assay of 

 bacteria washed from duplicate plates are shown as hollow dots. Those for 

 which growth was calculated from the average number of bacteria per micro- 

 colony are shown as solid dots. 



Something similar to the expected high early rate of appearance of resistant 

 clones is observed in these results. In figure 1 the numbers of resistant clones 

 appearing during growth are plotted on a logarithmic grid against growth ex- 

 pressed in terms of factor increase in numbers of bacteria. For comparison, a 

 curve is drawn showing the expected numbers — assuming a rate of .55X10 -8 

 and immediate phenotypic expression. The experimental curve does show a 

 high early rate, declining as growth proceeds. This is qualitatively what would 

 be expected where there is a delay, though the rate is actually higher than ex- 

 pected. 



The increase in resistant clones when material is sprayed with phage after 

 a twofold increase in population is approximately 12 per 10 8 bacteria plated, 

 which represents an apparent mutation rate of 8.3 X10 -8 ; whereas the rate 

 expected during the first division is the same as that obtained from liquid cul- 

 tures, that is, something between 2.8 and 3.4X10 -8 . 



Furthermore, by extrapolating the curve backward it would appear that 

 about seven per 10 8 bacteria become resistant during the lag phase before any 

 division has taken place. 



This point has been studied by incubating bacteria with a known lag phase 

 of 70 minutes for 30, 60, and 90 minutes on agar and then spraying with phage. 

 The results are given in table 7, and show an appreciable increase in resistant 

 colonies from plates sprayed just before the onset of the first division and dur- 

 ing the very early part of the division. 



The points just considered are not directly related to the main issue. The 

 important contribution of these experiments is to show that the rate of appear- 

 ance of resistant clones declines during the first few divisions on agar. 



67 



