Séance du 16 Avril 1924 
107 
lopment and, after reaching maturity, until it died, 
except in special cases when pairs were put together 
in larger cages for copulation. It was, therefore, quite 
impossible for copulation to take place except in these 
latter cases. 
Until the first moult, some of our specimens 
were bred on Oleander Aphis, Aphis nerii Boyr., as 
no flies small enough were available and the nature 
of our experiments did not allow us to employ the 
method previously used of allowing the newly- 
emerged larvae to struggle for survival : Adair 
(1914.1) p. 119; (1916.1) pp. 76-79; (1916.2) p. 9$. 
Nineteen out of ninety-seven of these in our first ge¬ 
neration died before the first moult, while only five 
out of fifty-seven of those which were fed on small 
Diptera and other minute insects caught in a beating 
net failed before that. In 1912, (1914.1) p. up, we 
had been unable to feed M. Savignyi on other species 
of Aphis, so that it is very probable that they do not 
normally eat these. Nevertheless we must not over¬ 
look the fact that those which did survive were as 
healthv as their sisters and that the time of vear was 
less favourable for their development, as they were 
the earliest. 
Later we used the rotten-fruit flv, Drosonhiia me- 
lanogaster Mg., of which large numbers can be ob¬ 
tained easily from rotten apples, tomatoes, stale stout 
or other similar substances throughout the year. 
After the first moult, they are strong enough lo 
tackle house flies and we used practically nothing 
else. Only seven of the t3o remaining after the first 
moult failed to reach maturitv, so that the diet mav 
