1948] 



ORIGIN OF BACTERIAL RESISTANCE TO ANTIBIOTICS 



07 



the same mixture of resistant and sensitive bacteria. Among the samples from 

 separate cultures, if mutations occur at random, a large number of resistant 

 colonies would be obtained from cultures in which mutation happened to occur 

 early in the growth of the culture, and a small number of resistant colonies from 

 cultures in which mutation happened to occur late, provided the growth rate of 

 resistant bacteria is not appreciably different from that of normal ones. If 

 resistance originates by mutation, then, the variation in numbers of resistant 

 bacteria would be much greater between samples taken from separate cultures 

 than between samples taken from the same culture. 



TABLE 1 



Number of bacteria (E. coli) resistant to a concentration of 5 units of streptomycin per ml of 



agar medium in samples taken from a series of independent cultures and similar samples 



taken from a single culture which assayed l.S X 10* bacteria per ml 



SAMPLES FROM INDEPENDENT CULTURES 



No. of resistant 

 bacteria 



56 

 91 

 123 

 97 

 48 

 52 

 54 

 89 

 111 

 164 



Average 105.9 



Variance 2913.9 



Chi-square 550.3 



P much less than 0.001 



SAMPLES FROM SINGLE CULTURE 



No. of resistant 

 bacteria 



110 

 125 

 135 

 121 

 112 



Average 131.2 



Variance 151 . 1 



Chi-square 17.3 



P 0.26 



Table 1 shows the results of such an experiment with E. coli and streptomycin. 

 In addition to the tests represented in the table, the concentration of bacteria was 

 determined in 11 cultures, including the one from which the 15 samples were 

 taken. The average number of bacteria in 10 cultures was 2.2 X 10 8 per ml, 

 with extreme variants of 1.9 and 2.3, and the average number in the eleventh was 

 2.1 X 10 8 per ml. Thus the variation in numbers of bacteria among the differ- 

 ent cultures was so small that it could have introduced only negligible differ- 

 ences between the numbers of resistant colonies observed on different plates. 

 It is evident from table 1 that the variation in number of resistant colonies was 

 considerably greater among platings from independent cultures than among 

 platings from a single culture. The extreme variants of independent cultures 

 were 48 and 291, the average 106, the variance 2,914, chi-square 550, and the 

 probability that this variation was due to chance is insignificant. On the other 



83 



