580 Notes in 1887 upon lepidopterous larve, éc. 
Further support is given by the small size of this 
species as found in Madeira and the Canaries. We may 
have the ancestral size preserved unchanged in these 
islands, or reversion may have taken place. The fact that 
other species do not seem to be similarly reduced in size 
is in favour of the former alternative. Careful investi- 
gation of the conditions under which the larve exist in 
these islands would doubtless show why they can there 
maintain the smaller size without danger, although they 
cannot do so in other places. In the meantime the 
following facts deserve attention:— the only insect- 
eating vertebrates in these islands are birds and lizards: 
in Madeira there is only one small species of lizard,—so 
small that the mature larva, pupa, and imago would 
certainly be sufficiently protected by the size which is 
usual in the island. Furthermore, these lizards, and the 
more numerous and larger species of the Canaries, have 
become largely frugivorous,—an almost certain result of 
the insufficiency of insect-food. It is therefore probable 
that such lizards, having learnt to eat other food, are 
not, like most of these reptiles in other places, inveterate 
enemies of all insect-life, although it is probable that they 
still eat a great many insects. The kestrels in these 
islands also eat large numbers of insects. I saw the crop 
of one of these birds which had been shot in Madeira: it 
contained hundreds of larve of some species of Noctua, 
a larva which is very common among the roots of the 
erass on the higher parts of the island. There was also 
a small lizard in the crop. The larve did not approach 
the size attained by S. convolvuli in the island. 
These conclusions may be met by the objection that 
the larvee must pass through all the smaller sizes in order 
to become mature. Of course this is true; but a larva 
which is protected by size, in addition to concealment, 
during part of its life, may gain advantage over a larva 
which depends upon the latter method alone for the 
whole of its life. Again, size would be especially ad- 
vantageous in the critical time during which the larva is 
wandering about before burying, and also in the imaginal, 
and, perhaps to a small extent, in the pupal stage. 
There are also some other interesting conclusions to 
be reached by a study of the diagram. 
The gradually increasing rate of growth in the 
successive stages is seen in the increasing steepness 
