TKAXSLOCATION I OF DROSOPHILA. 439 



TABLE: V. 



Cy ru h st p ss Tr Cy A H e Tr 



" V X rrr" ~~~~ Cj . 



e ro ca e ro ca 



The results given in the last three tables are summed up in 

 the chromosome maps shown in Fig. i, d, e, and/, and Fig. 2, b 

 shows the ratios of these values to the standard values. It will 

 be seen on comparing these results with those based on the 

 homozygous Tr flies that there is no noticeable difference in the 

 crossover effects produced by the homozygous and by the 

 heterozygous conditions. 



The chief significance of the results probably lies in their 

 bearing on the mode of attachment of the translocation. There 

 are two conceivable ways in which the translocated piece might 

 be attached between ebony and rough : it might be interpolated 

 within this length, thus causing a prolongation of the thread 

 between ebony and rough, or it might be attached somehow to 

 the side of the original thread, thus resulting in a configuration 

 of genes and of chromatin different from what had previously 

 been regarded as normal, and not entirely in one line. The 

 present results clearly disprove the first alternative, for if it were 

 true, crossing over between ebony and rough should be increased 

 when the flies were homozygous for the interpolated fragment. 

 We must therefore conclude that the chromosome carrying 

 translocation is branched or doubled in the region where the 

 latter is attached. 



The second point of significance is that the presence of trans- 

 location does not reduce crossing over throughout the chromo- 

 some. Beginning at the left-hand end of the chromosome, we 

 find that the number of crossovers is practically normal until 

 we reach the locus of pink, one third of the distance along the 

 chromosome. From this point on, the map distance is markedly 



