NATURE OF THE GENETIC EFFECTS 409 



the proliferation of the very cells in which they occurred, through their 

 effects on cell metabolism. Since the cells of the male contain but one 

 X chromosome there is no normal allele present to dominate over these 

 lethals. For this reason the mature spermatozoa that are finally pro- 

 duced, and that give rise to the offspring tested for mutations, sometimes 

 come to contain only a third as high a frequency of mutant genes as has 

 originally been produced. The amount of this germinal selection has 

 been gauged by Kossikov (1936, 1937) and by Serebrovskaya and 

 Shapiro (1935) by comparing the frequency of these mutations with the 

 frequencies obtained in the autosomes of males or in either the X chromo- 

 somes or autosomes of females when the same early stages in germ 

 cell history are irradiated. For the presence in these cases of a 

 homologous chromosome, bearing the dominant normal allele, renders 

 effects of the lethals on the metabolism of the cells containing them 

 negligible. 



Even when the error caused by germinal selection is avoided, by the 

 use of one of the methods just referred to, it is found that lethals are 

 induced by ionizing radiation in the gonial germ cells of Drosophila (repre- 

 senting for the most part ordinary interphase nuclei) with a distinctly 

 lower frequency than in spermatozoa. A similar, although probably 

 smaller, difference is found for specific visible gene mutations. In the 

 case of the lethals, it is still unsafe to estimate what portion of the excess 

 of the mutations in spermatozoa, as compared with earlier germ cells, is 

 to be ascribed to the higher rate of production of deletions and other 

 structural changes in them than in ordinary interphase nuclei. But, for 

 the specific visible mutations, laborious analyses of the individual muta- 

 tions have been carried out, in order to screen out the structural changes 

 as far as possible. The preliminary results of this work (Muller, R. M. 

 Valencia, and J. I. Valencia, 1950 and unpubUshed), which requires larger 

 scale prosecution, indicate that the residual class designated as gene 

 mutations is produced in ordinary interphase nuclei of germ cells (gonia) 

 with only about a half of their ordinary frequency in spermatozoa. 

 05cytes, at least, in their middle and later growth stages, and probably 

 spermatocytes also, have an induced frequency of gene mutations 

 more like that in spermatozoa than that in gonial stages (Berman, 1939; 

 Muller, R. M. Valencia, and J. I. Valencia, 1950).^ The results on 

 visible gene mutations produced in the somatic cells of embryonic or 

 young stages give frequencies of the same order of magnitude as obtained 

 with germ cells, but they are as yet too meager for a decision as to which 

 germ cell stage they resemble more closely in frequency. It is of course 

 to be expected that they would have frequencies like those in the gonial 

 (interphase) germ cells. 



3 The frequency in spermatozoa of different stages has recently been found to vary 

 considerably, as noted in Chap. 8. 



