304 CALVIN B. BRIDGES 



In the early experiments involving purple several other muta- 

 tions arose, probably the most interesting of which was 'extra 

 bristles,' which led to the study made by MacDowell on the 

 effect of selection on bristle number. 



In the attack upon the problem of ' inviability' purple entered 

 into the first experiment planned to include the balancing of 

 inviability by complementary crosses. This practice was ex- 

 tended to involve three locus experiments in the balancing of the 

 black purple vestigial back cross. The inviability of vestigial 

 met with in the early purple vestigial crosses seems not to have 

 been due primarily to ' prematuration/ 'repugnance,' or autoso- 

 mal lethals, but probably to culture conditions, as shown by 

 Carver (unpublished). 



In the development of autosomal linkage, purple was involved 

 in the first coupling F2, back-cross test of crossing over in the 

 male, and likewise in the female. One of these back-cross tests 

 of the male gave a few crossovers which prevented a clear con- 

 ception of no crossing over in the male. The back-cross tests 

 of the female gave the first Hwo-point' autosomal map, purple 

 vestigial. The first autosomal 'three-point' map was black 

 purple vestigial, completed by the determination of the black 

 purple crossover value. 



With that most fascinating and difficult subject — the analysis 

 of the relation between the physical chromosome and the process 

 of crossing over — purple has been intimately connected. The 

 relatively high coincidences obtained in the cases of black purple 

 curved and black purple vestigial soon showed that this relation- 

 ship in the purple region of the second chromosome is different 

 from the relationship for sections of like map distance in the 

 compared regions of the first chromosome. An expansion of 

 this comparative study should aid in arriving at the cause of the 

 differences. 



The change in coincidence accompanying the age variation in 

 crossing over in the case of black purple curved (Bridges, '15) 

 led to the tentative conclusion that both changes were mainly 

 due to a lengthening of the average length of the section of 

 chromosome between coincident crossovers, rather than to a 



