476 RADIATION BIOLOGY 



dose. At least, this relation has been found to hold, when spermatozoa 

 of Drosophila are irradiated, for the most studied class of mutations, that 

 of sex-linked lethals, a considerable majority (over four-fifths) of which 

 consist, at doses up to some 1500 r, of changes that are not connected with 

 any structural alteration readily detectable by an influence on crossing 

 over or by the appearance of the salivary chromosomes. As a sequel to 

 the preliminary but inconclusive work on the problem by Muller (1928b), 

 the first experiments on Drosophila which clearly showed this relation 

 were carried out at the University of Texas by Oliver (1930, 1932). The 

 doses used ranged from about 300-5000 r of X rays and were varied by 

 altering the time of exposure. In barley, Stadler (1928b, c) had already 

 reported preliminary results indicating this relation, and he gave further 

 evidence subsequently (Stadler, 1930, 1932). The principle has since 

 been extended to organisms widely different from Drosophila and barley, 

 such as the snapdragon Antirrhinum (Noethling and Stubbe, 1936), the 

 liverwort Sphaerocarpus (Knapp, 1935), and the fungus Neurospora 

 (Lindegren and Lindegren, 1941 ;Sansome, Demerec and HoUaender, 1945). 



In Drosophila the linear relation was soon confirmed by several investi- 

 gators and was shown to hold for visible mutations by Timofeeff-Res- 

 sovsky and Delbriick (1936), by Heptner and Demidova (1936), and by 

 Muller (1940). In the most recent work with lethals (Spencer and 

 Stern, 1948; Uphoff and Stern, 1949), the range of dose has been extended 

 down to 50 and 25 r. At these lowest doses, however, the relative error is 

 much larger, not only because of the minute total frequency of mutations 

 but more especially because of the relatively enormous variations (dis- 

 cussed in Chap. 7) to which the spontaneous frequency is subject and 

 which must be subtracted from the total for the derivation of the induced 

 frequency. It is not surprising therefore that, at these lowest rates (and 

 at these rates only), not all the experiments on sex-linked lethals gave 

 results in conformity with the principle of linearity. However, it should 

 be noted that the one experiment (Caspari and Stern, 1948) which failed 

 to do so gave an unusually high control frequency, the subtraction of 

 which from the experimental value resulted in the induced frequency as 

 calculated being too low. This experiment was at the same time more 

 doubtful than the others on some other grounds. All in all, the agree- 

 ment with linearity, found by numerous independently working investi- 

 gators, has been remarkably close, despite the skepticism of some non- 

 geneticists (e.g., R. D. Evans, 1949) who had suspected radiation of being 

 less efficient at low doses. 



Not only is the curve relating the frequency of induced sex-linked 

 lethals to dosage a straight line, but, as closely as a reasonable amount of 

 statistical error will allow, this line points directly to an origin at zero. 

 That is, when the same spontaneous-mutation frequency is subtracted 

 from each of the observed frequencies and a best-fit straight line is 



