388 H. G. MAY. 



It was impossible to find any noticeable effect of moisture, 

 consistency, or age of food on the facet number. No significant 

 difference was found in facet number between the first flies hatch- 

 ing from a bottle and the last ones except in cases where the 

 food gradually dried up and the last flies were minute, i. e., 

 one half the size of normal flies or less. Such minute flies showed 

 a tendency toward lower facet numbers. Accurate size measure- 

 ments and facet counts on a large number of flies may possibly 

 reveal a slight correlation between size of body and number of 

 facets, but mere observation failed to detect it. It is also 

 possible that a slight correlation exists but is concealed by other 

 factors that have more influence. 



No relation could be discovered between the age of the parent 

 and the facet number of the offspring. In case of such a relation 

 the latest offspring from two parents would be different from the 

 first. No such difference could be detected. 



Likewise it was impossible to find any definite correlation 

 between the length of the developmental period of the larva 

 and the facet number of the adult. No special experiments were 

 undertaken with material in which the exact developmental 

 period was ascertained, but, since in productive bottles the 

 period of hatching is much longer than the period during which 

 the eggs are laid, the first larva? must as a rule have a shorter 

 developmental period than the last, and any difference due to 

 i he length of the developmental period should become evident. 

 A glance at Fig. 8 shows that beyond the daily fluctuations 

 due to the fact that the number of flies examined was too small 

 to be representative, there are larger, parallel fluctuations of the 

 two lines. Since the food during the Ba selections was fairly 

 constant in character it can not be regarded as the cause of these 

 fluctuations. The temperature of the room in which the flies 

 were kept varied between rather wide limits. Unfortunately no 

 thermograph was available for the room at that time. There 

 is no correlation between the outside temperature and the facet 

 number; but the variations in the room temperature were inde- 

 pendent of those in the outside temperature. Since experi- 

 ments to determine the effect of temperature on the facet number 

 were undertaken by E. W. Seyster of this laboratory during the 



