312 Hubert P. Yockey 



irradiated diploid (••). This shielding seems to be complete for errors due 

 to first order damage; that is, damage such that J{X) = /„ in equation (6). 

 If this shielding were not complete a term in the first power of X would be 

 apparent in Fig. 6. 



Thus far the application of information theory to the currently accepted 

 model of the diploid cell succeeds very well. We find features we expect and 

 do not find ones we do not expect. Owen and Mortimer's data on the domi- 

 nant lethal survival enable one to study second order effects ordinarily sub- 

 merged in those of the first order discussed in the paragraphs above. 



Figure 6 shows that there is a very small dominant lethal expression of 

 damage but only when the damage is of the second order in the irradiated 

 parent. That is, a single error or group of errors is shielded but pairs of errors 

 or pairs of groups of errors are expressed, to some degree at least. This is 

 evident since log ///q behaves as P. It is to be expected that this higher order 

 damage exists in the haploid and in the diploid (••) but cannot be observed 

 because of the lethality due to first order damage. 



The survivorship has the same P behavior for the other ploidies, but a 

 curious feature is that this higher order damage is expressed to a greater degree 

 in the higher ploidies, contrary to expectation. Perhaps this is a model-sensitive 

 phenomenon (as higher order phenomena often are). If that is so further 

 experimentation may tell us more about polyploidy. 



It was pointed out in Section II that j{X) is related to the interaction of 

 radiation and matter. This indicates that repeating Owen and Mortimer's 

 experiments with other deleterious agents may be very fruitful. For example, 

 Uretz (15) has shown that the ultraviolet survivorship of haploid yeast is 

 sigmoidal. If this means in the case of haploid survival /(A) = J^X, these errors 

 will probably be shielded in the zygote. We choose the next higher term J{X) 

 = J^X^ so that log JJIq may be expected to behave as X^. Higher powers in X may 

 be found in the expansion of /(A) depending on the effectiveness of shielding of 

 recessive lethal mutations in the zygote. 



These results should apply to organisms other than yeast and in 

 particular to the survivorship of F^ progeny in mice. F^ progeny with one 

 irradiated parent should have a shorter life span than the unirradiated parents. 

 Fj progeny with two irradiated parents should have a still shorter life span. 

 That this is at least partly the case is shown by recently reported results by 

 Russell (34). He reports a life shortening in the offspring of male mice exposed 

 to neutron irradiation from a nuclear detonation. The dose was rather low; 

 the highest to the parent was 186 rep, but only two such oflfspring were obtained. 

 Rather small numbers of individuals were obtained from other parents also 

 so that the estimate of the magnitude of the effect is rough, although its existence 

 seems to be established. The life shortening seemed to be of the same order 

 of magnitude in the father as in the offspring, however. 



Wallace (62) has reported work on Drosophila in which he has irradiated 

 several populations for as many as 150 generations. His criterion of viability 

 is survival from egg to emergence and this work refers only to the second 

 chromosome. He finds that the fitness of a population does not necessarily 

 continue to decrease under the influence of radiation. 



These experiments together can be understood from the point of view 



