WHAT IS A GENE? 155 



as ad^ and 0^/23' ^^ closely linked nonallelic genes with clearly dif- 

 ferent functions such as yellow (y) and ad^ on chromosome I. The ad^ 

 and ad^ genes are clearly nonallelic on all three definitions of the gene. 

 It is otherwise with the seven adenineless mutants in the ad^ group on 

 chromosome I. They, like the ad^ and ad^ mutants, are nutritionally 

 indistinguishable, growing on the same compounds. None of the group, 

 however, will complement any other of its members, either in the hetero- 

 karyotic condition or when heterozygous diploids are formed (Chapter 

 12). Functionally thev behave as alleles, yet, when intercrosses are 

 made, adenine-independent nuclei arise during meiosis with frequencies 

 that cannot be accounted for by back mutation as it occurs in haploids. 

 These recombinations locate the seven mutational sites in sequence be- 

 tween the y and biotin {bi) markers which are only 6 map units apart. 

 The order is determined pairwise in this fashion: from the heterozygote 



y ad^ + 



there were 139 se.xual spores which ga\'e cultures sh(n\ ing 



+ ad^Q bi 



adenine independence; of them, 133 were yellow and six were green (Y ), 



suggesting the order y.ad^^-.ad^-.bi, in which case the composition of 



the heterozvgote should have been written ^ ^ — — . The 



-(- arfjg + bi 



segregation of the less closely linked bi confirmed this qualitative con- 

 clusion. When exchanges occur between these mutants, the reciprocal 

 recombination products may be derived from single nuclei (Chapter 12), 

 showing that the process meets the formal requirements of crossing over. 

 Considering for illustrative purposes interactions between just two of 

 these mutants, the results can be summarized as in Figure 6.2. It can 

 be seen that when either ad.^^ or ad^ is present alone in a heterokaryon 

 or a heterozygote, it is recessive and gives a normal wild-type phenotype. 

 The phenotype is normal whenever the two dominant factors are on the 

 same chromosome {cis)\ but when one dominant factor is on one chromo- 

 some and the other on its homologue (trans), the mutant character is 

 expressed. Thus ad^^ and ad^ behave as functional alleles in this in- 

 stance, but they seem to occupy two separable sites as do nonallelic 

 genes. The contrast can be brought out by considering the gene y which 

 is closer (0.048 units) to ad ^^ than the latter is to adg (0.14 units). Yet 

 where ad^^^ and ad^ behave as pseudoalleles, ad ^^^ and 1/ do not; the 



frans-heterozygote — ; — is wild type. 



+ ac/i6 



The bi factors at the adjacent locus give another example. All three 



will grow on 7,8-diamino pelargonie acid, desthiobiotin or biotin, but not 



on pimelic acid. 



