220 



presented which indicates that this 

 method makes possible the recovery of 

 mutants without regard to the nature 

 of the induced biochemical deficiency 

 or the composition of the (complete) 

 medium. 



Application of this method has 

 shown that the proportion of biochem- 

 ical mutants not recoverable by the 

 usual screening tests may be less than 

 50 per cent in Neurospora and less 

 than 25 per cent in E. coli. It is calcu- 

 lated that this rate of loss does not 

 produce a sufficiently intense selection 

 of unifunctional mutations to account 

 for the high frequency of such muta- 

 tions actually found. 



Finally, it is suggested that the re- 

 sults can be simply accounted for on 

 the hypothesis that the synthesis of a 

 protein molecule is accomplished by a 

 single catalyst working directly on the 

 constituent amino acids or their simple 

 derivatives. 



APPENDIX 



The corrected value of the fre- 

 quency of unifunctional genes can be 

 computed from a Poisson distribution 

 as follows: 



Assume that each gene has one func- 

 tion to begin with and that there is in 

 addition a number of functions, m, 

 randomly distributed among the genes. 

 The fraction i of all gene functions is 

 indispensable. Letting 



n = the number of genes 

 Then n -I- m = the number of gene 

 functions, 

 m/n = c = the mean number of 



additional functions 

 per gene, 

 And ic = the mean number of 

 indispensable addi- 

 tional functions per 

 gene. 



The fraction of genes with no addi- 

 tional functions— i.e., unifunctional— is 



HOROWITZ AND LEUPOLD 



then given by e^'', the first term of a 

 Poisson distribution. Of these, the 

 fraction (1— i) are recoverable. The 

 fraction of recoverable unifunctional 

 genes is therefore 



Pa= (l-i)e-<^ 



Similarly, the fraction of all genes, uni- 

 functional and multifunctional, which 

 can be recovered is 



Pa+b= (l-i)e-- 



The frequency of unifunctional genes 

 among those recovered is therefore 



Pa/Pa + b = e-^<l-'' 



Equating this expression to the ob- 

 served value, 0.84, and substituting 0.5 

 for i, one finds c = 0.34. The cor- 

 rected frequency of unifunctional 

 genes is then e"^^** = 0.71. 



REFERENCES 



Beadle, G. W. and Tatum, E. L. 1941 Ge- 

 netic control of biochemical reactions in 

 Neurospora. Proc. Nat. Acad. Sci. Wash. 

 27:499-506. 



Bonner, D., 1946a Further studies of mutant 

 strains of Neurospora requiring isoleucine 

 and valine. /. Biol. Che???. 166:545-554. 

 1946b Biochemical mutations in Neuro- 

 spora. Cold Spr. Harbor Syniposiin?! 

 Quant. Biol. ll:l+-24. 



Davis, B. D. 1948 Isolation of biochemically 

 deficient mutants of bacteria by penicillin. 

 /. Af/ier. Chem. Soc. 70:4267. 



1950 Studies on nutritionally deficient bac- 

 terial mutants isolated by means of penicil- 

 lin. Experientia 6:41-50. 



Doermann, A. H. 1944 A lysineless mutant of 

 Neurospora and its inhibition by arginine. 

 Arch. Biochevi. 5:373-383. 



Fmerson, S. 1950 Competitive reactions and 

 antagonisms in the biosynthesis of amino 

 acids bv Neurospora. Cold Spr. Harbor 

 Symposiuvi Quant. Biol. 14:40-48. 



Fincham, J. R. S., /. Gen. Microbiol, in press. 



