40 CELL HEREDITY 



fonvard or back mutants by manipulating the environment. Let us 

 consider, for example, mutants of the intestinal bacillus, Escherichia 

 coli, showing resistance to streptomycin. Even if they exist among 

 streptomycin-sensitive cells (str-s) in a frequency as small as one in a 

 billion (10~^), they may be found simply by spreading several billion 

 str-s cells on a plate of agar containing nutrients and streptomycin. Only 

 the str-r mutants will grow to form visible colonies; the str-s cells will 

 die. Among the resistant bacteria selected in such experiments, some — 

 streptomycin-dependent (str-d) — actually require streptomycin for 

 growth. When these str-d bacteria are spread on mutant agar without 

 streptomycin, only the rare str-s bacteria among them grow and form 

 colonies. These back mutants do not need streptomycin for growth and 

 are no longer resistant to it. They may be selected among billions of 

 str-d cells. In this way the mutation process is shown to be reversible. 



1 X 10-^ 

 str-s y — str-d 



1 X io-» 



The numbers along the arrows here indicate the rate of mutation. We 

 may consider the str-s state as the wild type even though in this case 

 it is the more stable. 



AN OPERATIONAL DEFINITION OF THE MUTABLE UNIT 



In the transformation experiments we saw that the genetic material 

 existed in mutually exclusive states, either str-s or str-r, and either 

 mtl^ or mtl~. These alternative states of individual genes are called 

 alleles. Members of the same allelic pair do not coexist in the same 

 bit of genetic material but can only replace one another. But the same 

 bit of genetic material contains individual members of different allelic 

 pairs so that the replacement of one of these by its allele in transforma- 

 tion results in recombination. We have also seen that, when the genetic 

 material is studied by the method of mutation, it is observed to pass 

 back and forth between alternative states. It is not surprising that the 

 fundamental genetic property of mutual exclusion of alleles should be 

 observed bv both methods since the alleles studied in transformation 

 initially arose by mutation. Clearly then, the genetic material consists of 

 mutable units which in their different states have different effects on the 

 phenotype. By what experimental operations can we distinguish one 

 mutable unit in its various allelic states from another mutable unit? 



