90 E. CARLETON MACDOWELL 



With these facts in mind it will be well to return to the gener- 

 ally low percentage of extras in the F2 of crosses with normals 

 of other races. It has previously been shown that apparent 

 normals may none-the-less be homozygous for extra bristles; 

 a plausible explanation of this fact seems to be the influence of 

 food. 



3. TEMPERATURE 



The regularity with which a new bottle produces an increase 

 in bristle numbers, if it follows a falling off at the end of the pre- 

 ceding bottle, shows that these variations are not mainly due to 

 temperature. This does not say that temperature may not 

 have a real, however slight, influence on bristle development. 



A series of preliminary experiments were performed to test 

 the influence of temperature. Sister matings were raised in 

 various temperatures, constant to a small fraction of a degree 

 contigrade. In one experiment different sets of eggs from the 

 same parents were raised in various temperatures and compared 

 with controls in room temperatures. Two main questions were 

 to be answered: Can the number of extra bristles be increased 

 by using higher or lower temperatures? Do constant tem- 

 peratures, irrespective of the degrees, influence the variability 

 of bristles and thus indicate an influence of temperature? In 

 the nine matings employed, 1860 flies were observed. The data 

 were all arranged in curves, males and females separated and 

 together, but no clear evidence was found to show that tempera- 

 ture had any influence, either in producing higher or lower num- 

 bers of extra bristles or by changing their variability. It is 

 very clear that temperature influences the bacterial and fungal 

 florae of the fermenting banana, and due to the chemical activity 

 of the food, the environment of the developing flies goes through 

 wide variations, as is witnessed by the multitude of difi'erent 

 shades of odors from a single bottle during its productivity. 

 Should any relation between temperature and bristle number 

 be found in later investigations, it would be extremely difficult 

 to show that this supposed relation was not an indirect effect, act- 



