412 EFFECT OF TEMPERATURE UPON FACET NUMBER 



Sources of Error. — Errors due to technique have been discussed at length by- 

 May. The chief sources are (1) errors in counting, and (2) dififerences in counts 

 between the right and left eyes. The latter factor is not significant in this study. 

 Counts were made on one eye only, but no preference was given either eye in mak- 

 ing the counts. The first factor was reduced to a minimum by using the low 

 facet stocks. Ultra-bar at 15° has a mean of only 60.81 facets for the males. In 

 such stock, all the facets are readily counted as no confusion of rows arises (Fig. 1). 



There is no direct correlation between the size of the fly and the number of 

 facets (Seyster, 1919). Exceptionally large individuals can be produced under 

 the best cultural conditions at 31° and yet show the extreme low facet number of 

 that temperature. All the flies are of practically the same size at all tempera- 

 tures. Fig. 1 shows the heads of two females, raised at 27 and 15° respectively. 

 They are drawn to exactly the same scale. 



Slight variations in the mean of the individual bottles and between separate 

 experiments may be attributed to slight differences in acidity, food, random samp- 

 ling, and other factors. 



Some of the variability in the Unselected stock may be due to germinal differ- 

 ences in the parents. The Ultra-bar and Low Selected bar stocks are practically 

 homogeneous. 



The precaution of keeping the parent stocks at a constant temperature was 

 unnecessary as shown by the experiments on the non-inheritance of the tempera- 

 ture effects. 



Bearing the foregoing points in mind, we may conclude that temperature is 

 the chief factor involved in the differences obtained in the mean facet numbers in 

 the present study. 



EXPERIMENTAL. 



Temperature Effect on the Mean Facet Number. 

 Temperature Effect on Unselected Bar Stock. 



In 1916, Seyster carried out some experiments on the bar-eyed 

 mutant of Drosophila showing that facet number varied inversely 

 with the temperature at which the flies developed. 



The preliminary experiments on Unselected bar stock confirm these 

 results. Summaries of the first experiments are given in Tables I and 

 II. They include the temperature at which the flies developed, the 

 number of individuals whose facets were counted, the average or mean 

 facet number for those individuals, the difference in the means ac- 

 companying a difference of 1°C. in the environment, and the per cent 

 of change in the mean produced by a difference of 1°C. 



