PROBLEMS OF MEASUREMENT OF MUTATION RATES 23 



and the calculations, but if you sediment and lyse, you're going to 

 produce ghosts. I'm curious to know just what strength of centrifuga- 

 tion you used. Would it have been sufficient to throw down the ghost 

 cells? 



Atwood: Yes, it would, and if you lyse chromium-labeled cells, the 

 ghosts contain from 15 to 20 per cent of the label. The remainder goes 

 into the supernatant and is mostly in the hemoglobin, probably because 

 the chromium binds generally to protein, and hemoglobin is the main 

 protein present. 



Lederberg: Do ghosts behave like the intact red cells in their agglu- 

 tination behavior? 



Cotterman: Yes, they do. 



Neel: Kim, I seem to recall that in other work where you presumably 

 were using a homozygous A individual, the proportion of exceptional 

 cells was higher than computed on the basis of two independent events. 

 Is that not correct? 



Atwood: Yes, that is correct, but it is not certain whether that 

 individual was homozygous A, although it is highly probable. Also, 

 the experiment was not carried enough stages to get a real level. It 

 began to hook over at about 10'^, so probably the level would not have 

 been much below that. Comparable A heterozygotes have about 10"^, 

 so there is a tremendous discrepancy between the expected (10"^) and 

 the observed (ca. 10~^) if that person is a homozygote. We now have 

 two known B homozygotes, and they show just this kind of discrep- 

 ancy. 



Before talking about these I want to mention two attempts to find 

 changes other than losses, that is, whether any B cells are produced 

 in A individuals. Experiments of this kind would not ordinarily be 

 meaningful because the B cells originating in an or an A individual 

 would not have a normal life span. The isoagglutinin would destroy the 

 circulating cells, and perhaps also the exceptional stem cells. 



However, we had a chance to see whether A goes to B in two 

 agammaglobulinemics who had no demonstrable anti-B, although they 

 were both A's. Judging from their non-A frequencies, they may be 

 heterozygous A. I say this with some reservation, because I am not 

 yet sure how often we would be wrong with this criterion. 



Cotterman: You had no family data on these two individuals? 



Atwood: Yes, but it is inconclusive. After the experiments, we tested 

 the survival of B cells in one of these agammaglobulinemics by injec- 

 tion of 2 ml. of Cr^i labeled cells. The survival of the injected B cells 



