212 



nary results which have been obtained 

 in our laboratory by Mr. Henry 

 Gershowitz, working with certain me- 

 thionine-requiring strains, indicate, 

 however, that the sterility can be over- 

 come by supplementing the medium 

 with a large quantity of the amino acid 

 —at least twice as much as is required 

 to produce optimal growth of vegeta- 

 tive cultures. This suggests that the 

 sterility may result from a high meta- 

 bolic requirement for the growth fac- 

 tor during the sexual process, or to a 

 lowered permeability to it. A fourth, 

 and relatively rare, residual effect is 

 failure of the mutant to attain a normal 

 growth rate. This also can characterize 

 an entire class of mutants, as in a cer- 

 tain group of strains of the cysteine- 

 methionine series now under investiga- 

 tion in our laboratory. 



Everything considered, it is perhaps 

 surprising that residual effects are not 

 observed more frequently, since in no 

 event is it possible, even in theory, to 

 avoid all of the consequences of a 

 mutation by supplying the deficient 

 metabolite. The block in the synthetic 

 pathway still remains, and it can have 

 an influence quite apart from the ef- 

 fects of the nutritional deficiency. It 

 has been shown in numerous instances 

 that metabolic intermediates may ac- 

 cumulate behind the block, sometimes 

 in spectacular quantities (for review, 

 see Horowitz, 1950). It would be sur- 

 prising if the presence of abnormal 

 concentrations of metabolic intermedi- 

 ates in the cells did not at times pro- 

 duce deleterious side-effects. Actually, 

 evidence has been obtained both in 

 Neurospora (Bonner, 1946a) and in 

 E. colt (Davis, 1950; Umbarger and 

 Mueller, 1951) that accumulated inter- 

 mediates may exert a lethal action by 

 interfering with reactions in other 

 metabolic pathways. This leads one to 



HOROWITZ AND LEUPOLD 



suspect that the so-called residual ef- 

 fects are to a large extent the irrepar- 

 able side-effects of the primary block. 

 (For further discussion, see Emerson, 

 1950.) 



THE SELECTION PROBLEM 



At the Cold Spring Harbor Sym- 

 posium of 1946, Delbriick raised a 

 question as to whether incompatibil- 

 ities with the one gene-one enzyme 

 hypothesis could be detected even if 

 they occurred (see Discussion follow- 

 ing paper by Bonner, 1946b). Del- 

 briick's argument was based on the 

 recognized fact that not all of the mu- 

 tations which are produced can be 

 detected by the methods usually em- 

 ployed for this purpose. Principally 

 three classes of biochemical mutants 

 are not recoverable: (1) those requir- 

 ing a substance which is absent from 

 the so-called "complete" medium used 

 for recovering nutritional mutants, 



(2) those requiring a substance which 

 is unable to diffuse into the cell, and 



(3) those requiring a substance which, 

 though present and diffusable, is not 

 utilized because of the inclusion in the 

 medium of an inhibitor of the mutant 

 in question. We shall refer to mutants 

 which, for the above, or for any other 

 reasons, are incapable of growing on 

 complete medium as mutants which 

 have lost an indispensable junction. 

 The point of Delbriick's argument 

 was that if any gene has more than one 

 primary function, it is likelv' that at 

 least one of these is an indispensable 

 function; in which case mutation of 

 the gene would not be detected. 



Now the validity of this argument 

 depends on the relative frequency of 

 indispensable functions. If this fre- 

 quency is very high, then the probabil- 

 ity of recovering a mutation of a 

 gene with several primar\' functions 

 will be very low. Thus, if 90 per cent 



