Notes in 1887 upon lepidopterous larve, dc. 558 
the other in the Canary Islands and in Madeira. In the 
larvee above-described it has been seen that three indi- 
viduals retained the green ground colour in the 4th stage, 
while all the others lost it. Under these circumstances, 
it might well have been supposed that the larve would 
exhibit the power of variable protective resemblance 
which so many species are now known to possess. Thus 
Rumia crategata can be rendered brown or greenish ; 
while greater or less variation in the depth of the brown 
ground colour can be caused in Crocallis elinguaria, 
Ennomos angularia, HE. lunaria, Boarmia rhomboidaria, 
B. roboraria, Catocala sponsa; and variations in the 
green colour of Smerinthus ocellatus and S. ligustri.* 
Other instances occur, and this power is doubtless very 
widely spread among larve. 
It is, however, certain that S. convolvuli does not 
possess this power in any marked degree, and it is 
probably entirely wanting. My larve were fed in 
glass cylinders placed upon white plates, and the food 
was always kept green and fresh. Only in the last 
stage were brown surroundings (earth and dead leaves) 
introduced, in order to test the opinion that the 
larve conceal themselves. The complete absence of 
brown surroundings in the earlier stages would have 
produced light-brown or green individuals of nearly all 
all the species above-mentioned, and yet the convolvuli 
larve became exceptionally dark. It is therefore probable 
that the predominance of dark varieties in one locality, 
and green in another, is due to the ordinary operation 
of natural selection upon the two forms of a dimorphic 
Species, tending to exterminate a relatively larger 
number of the variety which harmonises less with the 
surroundings in each locality. 
The regular food-plant of the species in this country and 
over the great part of Europe is Convolvulus arvensis (the 
larva being comparatively rare upon C. sepium). The 
* Professor Meldola was the first to bring together the scattered 
examples of this kind, and to draw attention to the general principle 
which they involve: see his paper, ‘On a certain class of variable 
protective colouring in insects’ (Zool. Soc. Proce., 1878, p. 153). 
Since then, many further details have been worked out: although 
some of the above-mentioned examples have not yet been published. 
See my papers in Proc. Roy. Soc., vol. xxxvili., p. 269, and xL., 
p- 135; also ‘Report of British Association,’ 1887, p. 756. The 
case of Boarmia rhomboidaria has been investigated by my friend 
and pupil, Mr. R. C. L. Perkins, of Jesus College, Oxford. 
