DIRECTION AND FEEQUENCY OF MUTATION 207 



adjacent ones. Each week the flies of the oldest bottle were 

 transferred to a bottle with new food and the old bottle was dis- 

 carded. The interval between the time that the flies were put 

 into a bottle and the time of removal was three weeks. The 

 length of the life-cycle at the chosen temperature is about four- 

 teen days with a minimum of twelve, so that there was no possi- 

 bility of the appearance of a second generation in any bottle. 



In the examination of the stocks for mutants the flies were 

 etherized and examined without counting the facets. The num- 

 ber of adult individuals was recorded. If no mutant was dis- 

 covered, the flies were put into a bottle with new food according 

 to the system given above. If, however, a mutant was discovered 

 all the flies were discarded and a new bottle was prepared from the 

 second bottle of the stock in case that did not have a mutant. 

 A selection was thus made against stocks exhibiting mutation. 

 There wals no object in doing this other than the desire to 

 preclude the possibihty of recording a single mutant more than 

 once. 



The only chance of carrying a mutation forward would have 

 come from a failure to recognize a mutant. In mutations from 

 full to bar or ultra-bar the differences in both males and heterozy- 

 gous females are too obvious to escape notice. In the reverse 

 mutations from bar to full the males offer no difficulties, but in 

 the unselected bar stocks heterozygous females coming from a low 

 bar individual might rarely be confused with high bar homozy- 

 gous ones. 



In the case of ultra-bar to bar the degree of dominance of ultra- 

 bar is so high that the heterozygotes are not readily distinguished 

 from ultra-bar by simple inspection without facet counts and 

 breeding tests. Here mutations might not be recognized until 

 the second generation, when they would be readily observed 

 because of the appearance of bar males. Male mutants are, 

 however, readily recognized at once. 



In the mutations from bar to ultra-bar there is usually no 

 difficulty in recognizing the heterozygotes because the latter are 

 closer to ultra-bar than to bar. Occasionally there may be a 

 doubtful case in unselected bar stocks where the presence of 



