Chromosomes and Genes 169 



completely new activity. To me both of these happenings after 

 duplication (first postulated by Bridges) appear to be rather mystic. 

 But I realize that some other geneticists have asserted that they 

 have no difficulty in envisaging this. 



We may at once oppose to this our OM'n interpretation of the 

 facts ( Goldschmidt, 1950c). It appears to be a necessary conse- 

 quence of the idea, which I regard as proved, of the position efiPect 

 segments, which we discussed at length. If a crossover break could 

 occur within such a segment — and there is no reason why it should 

 not — the described facts of pseudoallelism would be expected. Any 

 two "point mutants," deficiencies, or rearrangement breaks within a 

 segment are allelic, and if two of them, called mutants (which, we 

 know, behave like multiple alleles), are separated by crossing over 

 we get exactly the situation described for pseudoallelism. No position 

 alleles are involved, no position effect in the old sense of Sturtevant, 

 and no pseudoallelism, just ordinary allelism within the new concept 

 of the genie material. 



After this general statement of fundamentals, let us consider 



some of the detailed facts. A number of cases have been described 



by Lewis (see 1951) which fit exactly into the general description 



given for two "pseudoalleles." Such are the mutants Star and asteroid, 



Stubble and stubbloid, white and apricot. More remarkable still are 



those involving three mutants. There is first the work of Green and 



Green (1949) on the lozenge locus. Among many lozenge alleles, all 



of which affect the quantity of eye pigment, the eye structure, and 



the absence of spermathecae in the female, and all of which behave 



as a typical series of multiple alleles, three groups could be shown 



to exhibit a small amount of crossing over (resulting in one normal 



chromosome and one chromosome with more than one lozenge locus ) . 



Thus they behave like individual loci in close proximity, but permit 



crossover breaks between the pairs of alleles. Nevertheless, they act 



as alleles. This is best realized if combinations with at least one 



allele in both chromosomes are compared with those in which one, 



two, or three alleles are present in only one chromosome. In a 



standard case of Mendelian inheritance, individuals should all be 



a+c ab+ abc 



normal in these constitutions: — -, , ; — > or - — ; — -• Only the last 



+b+++c +++ 



is normal, however; the others show the compound lozenge effect: 

 a, b, and c behave as pseudoalleles. This disagreement with the 

 elementary rule of genetics is explained by the assumption that 

 a, b, and c, if located in different chromosomes ("trans," according 



