MUTAGENS OF POTENTIAL SIGNIFICANCE 229 



can't conceive of your collecting an adequate sample to get an ap- 

 preciable result. Do you really think you could? 



Spuhler: The sample could contain several thousand per year. 



Lederberg: But suppose all the spontaneous mutation rates in man 

 were due to caffeine; do you still think you would be able to pick 

 it up on the basis of this type of questioning? 



Atwood: Certainly; it is a matter of common experience that coffee 

 drinkers are fertile. An individual who is sterilized from caffeine must 

 take an awful lot more of it than most people do. You would be able 

 to tell by asking him. 



Magni: There was a positive correlation between the age of the 

 woman's father (at the birth of the woman) and the frequency of 

 women with zero children in Cavalli's data. A sample of the size used 

 by Italian Central Statistical Institute, as mentioned before, might 

 perhaps be adequate for testing the problem. 



Neel: Dr. Goldstein has reproached me for rejecting out of hand 

 the possibility of studies on human populations, and perhaps I was 

 overhasty, but at the moment I can't visualize any situation involving 

 two groups strictly comparable in all the ways that you would like 

 to have them comparable — one group ingesting a large amount of 

 coffee and the other not. Perhaps somebody else can think of such a 

 situation. 



Magni: Are you asking whether it would be possible to find two 

 identical groups? 



Neel: As nearly identical as is possible, except in this factor of 

 coffee consumption. 



Lederberg: I don't see that you could do better than the Cavalli 

 system in finding what heterogeneity there is in the population. It 

 may be stratified in other respects, but the particular beauty of it is 

 that you have been comparing not just two populations, but you are 

 comparing the age dependence of the sex ratio, and it would at least 

 give you some assurance that the difference you are speaking of is 

 a difference in mutation rate. If you found a difference, you would 

 then face the problem of whether it was the caffeine that differentiated 

 these populations or whether it was something else that was asso- 

 ciated, in some indirect way, with caffeine consumption. But you 

 would be on much stronger grounds in discussing this in relationship 

 to mutation. 



Neel: I suspect that coffee consumption has strong socio-economic 

 implications. 



Magni: This can be taken into account because the questionnaire 



