174 Nature of the Genetic Material 



inactivates the gene. In some cases recombination between alleles 

 will be possible. The phenotype of the double heterozygote in re- 

 pulsion is then as expected, since, in this diploid, both biotin genes 

 are inactivated by mutations at different sites. Assembly of all the 

 normal parts of the gene on a single chromosome is necessary for the 

 normal functioning of the gene." Needless to say, Pontecorvo does 

 not use the term "gene" in the classic sense, and thus agrees with 

 my conclusions at many points. 



Recently Green (1954) has studied a new case of pseudoallelism 

 at the vermilion locus. He finds that the two "genes" respond differently 

 to the action of specific suppressors; this indicates to him that the 

 pseudoalleles are really two different genes. This conclusion pre- 

 supposes that multiple alleles cannot show such different relations 

 to suppressor action. But I have shown in another case (the svr""* 

 alleles; Goldschmidt, 1945a) that within an allelic series different 

 alleles may have different qualities as suppressors. If this is possible, 

 there is no reason why the reaction to suppressors (as opposed to 

 reaction of suppressors) should not differ among alleles. 



It seems that in microorganisms much additional evidence of 

 the same type is available. Kaplan (1952) has reviewed a consider- 

 able series. Besides the mutants in Aspergillus and Neurospora al- 

 ready cited, he emphasizes the streptomycin-resistant mutants of 

 Escherischia coli (Demerec et ah, 1949&, 1950fl). It seems that there 

 is a rather large series of closely associated loci for different degrees of 

 resistance, concentrated in a small region of the chromosome map. 

 Kaplan accepts for these facts my interpretation in terms of a 

 chromosomal segment, which he calls "isophenic segments" as op- 

 posed to "genes," a term which might be adopted. 



In the present discussion we have emphasized cases in which a 

 chromosomal segment was involved in different steps of chemical 

 synthesis, though formerly we discussed pseudoalleles which showed 

 definite morphological effects. It may be added at this point that 

 cases exist in which it must be supposed that the causes for different 

 but closely related morphological effects are located within such a 

 segment. (The facts will be discussed also in a later reference to 

 chromosome sections.) We have called attention to the work of 

 Dunn, Gluecksohn-Schoenheimer, and others on tail mutations in the 

 mouse. A series has been discovered in which the members are all 

 closely linked, though a crossover analysis is rather difficult. A 

 lucid discussion of the implications has been presented by Glueck- 

 sohn-Schoenheimer (1949a) and Gluecksohn-Waelsch (1951). Gener- 



