14 



THE MATHEMATICAL TREATMENT OF DATA 



Dilute starting culture 

 of bacteria 



Growth period of several hours 



Concentrated final culture 



Concentrated final 



culture distributed 



1 ml per tube into 



many tubes 



Fig. 5. Bacteria are grown for a given period of time from a dilute starting 

 culture to a concentrated final culture. This final culture is distributed in 1 ml 

 aliquots into many tubes. These tubes are then sampled to determine the num- 

 ber of mutant bacteria in each tube. 



unity to warrant the deduction that there are different mechanisms in- 

 volved in producing the fluctuations which result in the deviations. As 

 with the deviation ratio test, the actual conclusion is that by chance 

 alone a ratio this large or larger will occur in a certain fraction of cases 

 with the given number of independent measurements in each of the vari- 

 ances. And, once again, the question of significance is decided on the 

 basis of one chance in 20 being significant and one chance in 100 being 

 highly significant. 



As an example, we take an investigation which was historically a 

 turning point in the understanding of the mechanism of mutations. The 

 appearance of a mutant cell is detected by placing the cells in an environ- 

 ment in which the mutant property can be expressed. An illustration from 

 bacterial genetics would be the appearance of resistance to a given anti- 



