162 CELL HEREDITY 



linked in the same order as the reaetion steps with whieh they are con- 

 cerned, and yet each unit is further subdivisible by mutation and recom- 

 bination. Some recombinable genes cannot transduce one another and 

 therefore seem to reside at the same site. A different situation is found 

 with mutants such as 22, which cannot transduce other members of its 

 group, although thev can transduce one another. Mutant 22 seems, 

 therefore, to have a location that covers the other members of group B. 

 It could be a deletion. The same reasoning has been applied to the other 

 mutants shown as bars rather than points in Figure 6.4. 



Although the bacterial cell is haploid, it is possible to test the be- 

 havior of the mutable units in cis- and frans-configuration by using 

 abortive transduction. A majority of the transductions do not involve 

 incorporation of the genetic material of the donor into the reproducing 

 genetic system of the recipient. The transduced fragment exerts its in- 

 fluence on the host's phenotype but is passed to only one of its progeny. 

 This constitutes a line, and on minimal medium a tiny colony is the 

 result (see Figure 5.12). Streaking such small colonies on minimal agar 

 results in the formation of only one colony, that formed from the cell 

 which inherited the transduced fragment. Whenever a transduction ex- 

 periment is carried out between mutants in different segments, abortive 



transductions result, e.g., — . Since the nonmutated seg- 



^ + hisD-l8 



ments complement one another, they behave as independent functional 

 units; they meet the cis-trans test for allelism. It is otherwise when 

 transductions are performed between different mutants in the same seg- 

 ment. There abortive transductions fail to occur since frans-arrange- 



ments within a functional unit, such as — — fail to comple- 



+ hisD-l8 



ment each other. On this basis they appear as different defects in the 



same functional unit, neither of which, as a consequence, can enable the 



formation of histidine from histidinol. The functional gene is clearly 



subdivisible into smaller mutable and recombinable units. 



The discovery that in Salmonella loci concerned with sequential steps 



in a biosynthesis are arranged in that order on the chromosome came as 



a surprise, because such linkages had not been found in Aspergillus and 



Neurospora. We can understand the finding in the following way. 



Neighboring genes may arise as duplications or repeats of a part of the 



chromosome. One or more of these repeats may gradually differentiate 



to assume a function related to but not identical with the original one, 



becoming thereby functionally nonallelic to its neighbor. This idea has 



much theoretical attraction, including the observation that the product 



of one gene is the substrate of its neighbor. These differentiated dupli- 



