Chromosomes and Genes 173 



just described which showed an extreme bx phenotype as position 

 effect, as expected within our interpretation. 



The question arises now whether some further argument for 

 one or the other genetic interpretation can be derived from a study 

 of the different phenotypic effects of the "pseudoalleles." This prob- 

 lem was first put by Pontecorvo (I952a,b); see also Roper (1950; 

 and in Pontecorvo, 1952, 1953). We have mentioned that they found 

 cases resembling those of pseudoallelism in Drosophila in the 

 mold Aspergillus involving synthesis of biotin and para-aminobenzoic 

 acid. The experiments were performed upon the hypothesis that 

 "the action of genes is in some way mediated by corresponding 

 enzymes, either each gene playing a part in the synthesis of each 

 enzyme, or on the conditions for the enzyme to act, including struc- 

 tural organization." Should these enzymes, by acting on a series of 

 reactions of millimicromolar order, be arranged in proximity to one 

 another, vestiges of this spatial organization might be reflected in 

 the corresponding genes. Thus the positive result for two cases of 

 mutants affecting different steps of synthesis of one substance made 

 Pontecorvo think (with emphasized caution) that "one of the ways 

 in which genes are clustered along the chromosomes is that of small 

 assembly lines of those genes which are involved in successive steps 

 of the metabolism of a scarce, ephemeral or non-movable intermedi- 

 ate." 



Pontecorvo realizes that the facts may be interpreted in two 

 different ways (which are those of Lewis and myself), and in his 

 latest review (1953) draws the following conclusions from his ex- 

 tensive work with the biotin alleles: "Two explanations at present 

 seem possible. The first, and less likely, is that the results are con- 

 validating the working hypothesis which prompted the investigation. 

 The wild type alleles at the three bi loci would then control different 

 biochemical reactions, the differences not being detected by the tests 

 used. The wild type alleles would function normally only when they 

 were all three on one and the same chromosome. This would be the 

 case if the intermediates in the reactions, controlled by the wild 

 type alleles were labile, or nondiffusible, or present in very low 

 concentration. Thus the unexpected phenotypes would be explained." 

 The essentials of the second and more plausible hypothesis are that 

 "the gene as a working unit in physiological action is based on a 

 chromosome segment larger than either the unit of mutation or 

 recombination. Mutation at different sites in the biotin gene gives 

 at least three, and possibly many, mutant alleles, any one of which 



