MANNER OF PRODUCTION OF MUTATIONS 591 



infertility. Viridis, however, although connected with infertility, was 

 not markedly increased in frequency by colchicine. Hence these findings 

 call for further genetic analysis. Using nitrogen mustard on barley seed, 

 Gustafsson and MacKey (1948) found that, in contrast to the results 

 from X rays, viridis was produced far oftener than albina. 



In a series of studies on visible mutants produced by irradiation of 

 Antirrhinum, Kaplan (1939, 1940a, 1946, 1948c), using pollen, and Knapp 

 and Kaplan (1942), using seeds, found apparent differences in the rela- 

 tive frequencies of different phenotypically defined classes of mutations, 

 depending on whether the material was irradiated in the dry or soaked 

 condition. When pollen was used, the R group of mutations, expressed 

 either as changes in (not absence of) chlorophyll or as absence of antho- 

 cyanin, appeared to be little if at all increased in frecjuency by soaking 

 the pollen before irradiation, whereas the Q group, expressed either as 

 absence of chlorophyll or as form and growth changes, was increased con- 

 siderably. However, the error for the R group was considerably greater 

 because of its unusually high and variable control rate. Curiously, the 

 division into categories, according to the effect of soaking on frecjuency, 

 appeared to cut along a very different line when seeds were used for 

 irradiation. For in that case, the abnormalities which seemed to pre- 

 dominate when dry seeds were irradiated were those visible in young 

 seedlings, and when wet seeds were used they were those visible in later 

 stages. Kaplan warns, however, that both the viabilities and classi- 

 fiabilities of different types vary greatly from plot to plot, according to 

 environmental conditions. It is not unlikely that this would also be 

 found true in barley. 



With barley seeds, Kaplan (1950c, 1951), although finding considerable 

 differences in the relative frequencies of visible mutations and chromo- 

 some changes according to the state of hydration and accompanying 

 chemicals (see Sects. 15 and 17), nevertheless failed to find significant 

 differences in the ratios of different phenotypic classes of visible mutations. 

 Moreover, unlike Gustafsson, he failed to find that certain classes are 

 more correlated than others with infertility of the plant arising from the 

 irradiated seed. 



In E. coli strain B/r, Newcombe (1952) showed conclusively in 1949 

 that there were pronounced differences in the ratios of different kinds of 

 mutations to streptomycin resistance and of those to phage Tl resistance, 

 according to whether the mutations were spontaneous or induced by ion- 

 izing radiation or by ultraviolet. The three types into which he classified 

 streptomycin resistance were sr (merely resistant), sdo (partially depend- 

 ent on streptomycin), and sdx (completely dependent). Among the 

 spontaneous and y- and ultraviolet-induced mutations these three types, 

 taken in the foregoing order, occurred in approximately the ratios 3:4:1, 

 2:2:1, and 1:1:2. Moreover, differences appeared when the sdo type 



