156 



Furthermore, they do not cross-feed one another, nor will stocks carrying 



one bi gene form biotin-independent heterokaryons with stocks carrying 



another. In this way the bi factors behave as functional alleles of one 



another. They are, however, separable as is shown by the fact that in 



hi /A -A- 

 crosses through the trans-heterozygote p^ they recombine to give 



the biA^ biB^ biotin-independent type which can be selected for on 

 minimal medium. 



The frequencies of recombinants indicate that they are arranged in 

 this order between the y and the ad^ loci: 



y bi2y bis ybi^ ^ad^ 



5.9 



0.1 0.4 0.7 



The behavior of the y and ad^ markers in crosses yielding biJ^ biB^ 

 types is compatible with their exact locations on chromosome I as deter- 

 mined in other crosses. This indicates that the biA^ biB^ recombinants 

 arise by a process of crossing over. 



It is their behavior in heterokaryons and in heterozygous diploid nuclei 

 that differentiates the bi factors in behavior from ad^ and ad.^, and 

 makes them resemble the factors in the ad^ group. In frarw-con figura- 

 tion the mutant phenotype is expressed. They behave, therefore, as 

 pseudoalleles. 



There are two reasonable explanations of this phenomenon. The first 

 presumes that chromosome I has a segment 0.1 map unit long which 

 functions at some step in the conversion of pimelic acid. This segment 

 has at least three mutable components which can be separated by cross- 



