PROBLEMS OF MEASUREMENT OF MUTATION RATES 13 



marrow cells, which is about 100 days. This may or may not be the 

 same for testis. 



Magni: Do you assume that the mutation rate is proportional to 

 the number of generations regardless of the generation time in differ- 

 ent stages of development? 



Atwood: The question is, which should you assume? The estimate 

 I just gave is based on the notion that the mutation rate per cell divi- 

 sion is the same whether the cells are dividing rapidly or slowly. 



Magni: So, you believe that mutation rate is constant for division 

 and not for absolute time? 



Atwood: Yes, that was the case just considered. If you assume, on 

 the other hand, that it is absolute time, then we can ignore the embry- 

 onic period, for the mutant accumulation would be linear with time. 



Lederberg: Aren't there more direct ways of getting the turnover 

 rate of the testis? 



Atwood: One way is by means of a pulse of P^- (28). 



Lederberg: Are those data known? 



Atwood: They are known for the rat and the mouse, but not, to 

 my knowledge, for man (27) . 



Lederberg: What about the rate of production of spermatozoa 

 divided by the number of spermatocytes? Or the total number of 

 gonial cells in the testis would be more appropriate. 



Atwood: You're asking how could you estimate the division rate 

 of the spermatogonial cells from the rate of production of the sperm, 

 as observed, and the number of stem cells present? I'm not sure what 

 the rate of production is. 



Lederberg: I suppose these are data that one could get, particularly 

 for man. They must be fairly easily available. 



Atwood: Yes. We have considered two alternatives: one, that the 

 rate per division stays the same and, in this case, the mutant ac- 

 cumulation shows a step increase followed by a gradual increase. Then, 

 that the rate per time unit is the same, in which case the early period 

 does not contribute disproportionately to the mutants. We could 

 also suggest that, perhaps, the mutation rate is dependent on the 

 rate of division; in other words, that the mutation rate per division 

 is higher for fast dividing cells than for slow dividing cells. In this 

 case, essentially all the mutants are added at the beginning when the 

 division rate is fast, and then you would notice hardly any effect of 

 age on mutant frequencies since most of the mutant cells would be 

 already there at birth. 



If we consider the age dependence of human mutation rates we can 



