G. S. Graham-Smith 
461 
The first generation laid eggs from July 12 to Aug. 6. From the 
first of these eggs flies emerged in 29 days. About 500 flies of the first 
generation were still surviving when members of the second generation 
began to appear on Aug. 10. This generation continued to emerge until 
Sept. 5, 3562 flies belonging to it being recorded. Owing to the adults 
belonging to two generations being intermingled it was impossible after 
this to be certain as to the exact numbers belonging to the succeeding 
generations, but the third generation began to emerge on Sept. 6 and 
continued to do so up to Oct. 2. About 4056 individuals belonging to 
this generation were counted. 
Towards the end of the season 2542 flies belonging to the fourth gene¬ 
ration appeared, and probably the majority of the pupae which hibernated 
in the ground through the winter of 1915-16 belonged to this generation. 
The descendants of the 100 females, which emerged from winter pupae, 
numbered during the season about 13,000 adults, instead of the 
101,250,000,000 postulated in four generations by the calculation 
employed by Howard, or to put the matter in figures which can be 
more readily understood, the descendants of each female during the 
season numbered 130 instead of 1012 mi llions. 
Chart 2 shows the number of living flies present on each day of 
the season, and the number which emerged on each day. It indicates 
that the flies which emerged from the winter pupae were the most 
hardy, for the early mortality amongst them was slight and they 
diminished in number gradually. In the subsequent generations, 
however, many of the flies only lived a few days, and probably never 
attained sexual maturity. Had these specimens lived longer and 
deposited eggs the columns representing the numbers present would 
probably have been several times higher. 
It may be argued that this mortality amongst the young flies was 
due to some defect in the experimental conditions, since it does not 
occur amongst flies reared in captivity indoors. The conditions pre¬ 
vailing in such experiments are not comparable, however, to those 
which prevail out of doors and in natural conditions, since the effects 
of rain, moisture, wind, low and high temperatures, lack of food, 
competing larvae and other enemies are eliminated. No doubt the 
results would have been more convincing had it been possible to 
conduct two similar experiments. Such a scheme was contemplated 
but abandoned owing to the labour involved. To some extent however 
a control is provided by the numbers caught daily in the various traps 
(p. 483). 
30—2 
