PROBLEMS OF MEASUREMENT 

 OF MUTATION RATES 



KIMBALL C. ATWOOD 



Department oj Microbiology 

 University oj Illinois, Urbana, Illinois 



Mutations, as everybody knows, are changes in nucleotide sequence. 

 The spontaneous rates of such changes have been estimated in differ- 

 ent organisms with varying degrees of indirectness. I would like to 

 compare the situation w'ith respect to such estimates in man and in 

 simpler organisms. 



This graph (Fig. 1) shows two distributions of reported mutation 

 rates. The group over here (open bars) is in man, and the group here 

 (solid bars) in microorganisms. The most frequent reported rate in man 

 is between 10"^ and 10"^. 



Freese: Per what? 



Atwood: What it is per is really not clear, as you will see, but it is 

 commonly expressed per gamete, or per human generation. In micro- 

 organisms the usual rate is somewhere between 10"^ and 10"^ either per 

 generation or convenient time unit, which is usually some time not 

 very different from generation time or doubling time. 



Auerbach: Per locus? 



Atwood: Whether it is per locus is not known certainly for all cases. 

 I have included mistakes that may have been made, as well as any 

 valid rates that happen to be in the group. 



Demerec: How did you get these data? Are they all from the litera- 

 ture? 



Atwood: Yes. There are about 50 estimates in microorganisms, about 

 half of which were reported by the Carnegie group. 



Glass: Are these per cistron, or per microbial group, or per site or 

 what? 



Atwood: They are mixed. I have not excluded any categories except 

 those where a mutator gene was known to be involved or in which 

 there were clearly labile alternative states, such as phase variation 

 and things of that sort, that tend to have high rates. 



