252 



of the Dt response provides a criterion 

 to exclude gene loss in the interpreta- 

 tion of experiments on spontaneous 

 and induced mutation of A. A tech- 

 nical advantage of a different sort is 

 provided by the R alleles. The pheno- 

 typic effect of R is such that a large 

 number of alleles may be objectively 

 distinguished by very slight differ- 

 ences of plant color intensity and pat- 

 tern. A gene with equally variable 

 allelic forms, if identified only by its 

 effect on some all-or-none response, 

 would seem to have only two alleles, 

 and its mutations would not be detec- 

 table except for those that crossed the 

 line between these two distinguishable 

 levels of action. Another advantage of 

 great practical importance is that both 

 R and A are genes affecting endosperm 

 characters and are, therefore, suitable 

 for the identification of mutations in 

 large populations. Both are apparently 

 genes of such trivial effect physiologic- 

 ally that their mutants survive with no 

 detectable loss of viability. 



The effective analysis of the diverse 

 genetic phenomena that may result 

 in the origin of a Mendelizing varia- 

 tion may not be impossible in intensive 

 studies of the mutations of suitable se- 

 lected genes, despite the fact that it 

 seems hopeless in studies of mutation 

 at miscellaneous, unspecified loci. 



These considerations are of no ac- 

 count if the frequency of spontaneous 

 mutation of the single gene is actually 

 too low to permit effective experi- 

 mental study. We cannot safely avoid 

 this difficulty by selecting for study 

 the genes of unusually high mutation 

 frequency, because there is no assur- 

 ance that the mechanism responsible 

 for the behavior of "unstable genes" is 

 representative of the mechanisms con- 

 cerned in typical gene mutation. The 

 use of microorganisms that permit ef- 

 fective screening for mutants in virtu- 



STADLER 



ally unlimited populations would re- 

 move the difficulty, but unfortunately 

 these do not provide the critical gene- 

 tic background essential to the study. 



A technique for determining the 

 spontaneous frequency of mutation of 

 specific genes is practicable in maize 

 for mutation rates ranging as low as 

 about one per 1 million gametes (18). 

 A test of eight genes, unselected ex- 

 cept for the technical advantage of 

 showing their effects in the endo- 

 sperm, yielded mutations in all but one 

 of the genes tested, the mutation fre- 

 quencies ranging from about one to 

 about 500 per 1 million gametes tested 

 (19). The genes that yielded mutations 

 in sufficient numbers to permit the 

 comparisons showed rather wide varia- 

 tion in mutation frequency in different 

 cultures. The gene R, for example, 

 yielded no mutations in large popula- 

 tions in some cultures, but its muta- 

 tion rate in other cultures ranged as 

 high as 0.2 percent. Later studies have 

 shown that such differences are due 

 in part to differences intrinsic to the 

 R allele concerned and in part to dif- 

 ferences caused by factors modifying 

 the mutation rate of R (20). Such fac- 

 tors are apparently quite common, 

 since a study in which only strong ef- 

 fects could be detected indicated the 

 occurrence of such modifiers in three 

 of the seven regions marked (21). 



The average mutation rates deter- 

 mined are rather low for effective ex- 

 perimental investigation of factors af- 

 fecting the mutation rate and even for 

 the extraction of adequate samples of 

 mutants for individual study. How- 

 ever, the fact that mutation rates are 

 so readily affected by diverse modi- 

 fiers makes it feasible to extract 

 strains in which the mutations of spe- 

 cific genes may be made frequent 

 enough to permit direct experimental 

 study. 



