PROBLEMS OF MEASUREMENT OF MUTATION RATES 17 



We have evidence that, in yeast, mutation rate in meiosis is higher than 

 in mitosis by a factor of ten for the specific mutons analyzed. If this 

 holds true also for man, one should expect two consequences: first, an 

 estimate by excess of the amount of mutations accumulated during 

 the embryonic period and, second, an apparent decrease or even a dis- 

 appearance of the age effect. 



Atwood: If one wanted to complete this notion of different periods, 

 I suppose one would have to separate off the stage just following the 

 embryonic period up to when some sperm are produced, as a sort of 

 static condition in which there is less division than during active 

 gametogenesis. 



I want to work up to the present state now, of a study that was 

 begun a few years ago, too optimistically, for the purpose of getting 

 human mutation rates from cells instead of people. The expectations 

 were that if we could find a method of getting the frequencies of 

 mutant cells in individuals, then we could deduce mutation rates from 

 these frequencies and some assumptions about the cell population 

 dynamics. 



For this, the ABO locus was chosen, because reagents and methods 

 suggested themselves immediately. I should point out that even if 

 germinal mutation rates at the ABO locus are relatively high, this 

 would be hard to detect. The ABO genotype frequencies are meaning- 

 less with respect to determination of mutation rates, because they 

 represent a balanced polymorphic system. We don't know what the 

 selective forces are that keep this system in balance, but they must be 

 potent ones, and capable of operating in different genetic backgrounds. 

 Both man and apes are still polymorphic for the ABO system, despite 

 their other genetic differences. 



One could attempt to get such mutation rates from aberrant pedi- 

 grees; for example, where an AB mother has an child, or an 

 mother has an AB child. These are the only instances in which an 

 ostensible mutation would be generally accepted; other cases in which 

 the identification of the mutation was dependent on the paternity of 

 the child would be equivocal, the mutation rates being lower as a rule 

 than illegitimacy rates.* 



We started to look at blood, then, with a method designed 

 to detect low frequencies of non-A or non-B cells in circulating eryth- 



* Steinberg estimated the probability that a mutant gamete would find itself in 

 a detectable zygote as : A or B — > O .02 



— > A .04 



O — >B .18 



