LEPIDOPTEROUS PUP2E AND THEIR SURROUNDING SURFACES. 
407 
were selected August 24, and divided into two lots of four each, so that the sizes of 
the larvge in the two lots were as equal as possible. Each lot was placed in a clear 
glass cylinder (of about 8 centimetres in internal diameter and I‘8 decimetre in 
height) with a roof and floor of brown paper. The food-plant (fennel) was confined 
to the lowest part of the cylinder and was very small in amount, and it was removed 
directly the larvge ceased feeding, while the rest of the cylinder was completely filled 
with dead and brown twigs of a coniferous shrub and the dry brown stems of 
Aconitum napellus with the brown seed-vessels adherent. 
The four larvge in one cylinder were carefully blinded August 24, a process which 
could be very effectually carried out in this species, for the larvge are very quiet, and, 
furthermore, all the ocelli are placed upon a distinct black patch ; and when the latter 
is covered with an opaque varnish the ocelli must be also covered. On August 25 the 
larvae were painted with varnish a second time, and on the 26th a third time. The 
other larvge remained normal. 
August 26, 9.20 p.m. One of the blinded larvge had changed into a green pupa on 
the brown floor, but in close proximity to the green leaves of the food-plant, and one 
of the normal larvge had also changed into a green pupa, which was fixed to one of 
the brown coniferous twigs ; but here also the green food-plant was just beneath. The 
food-plant was then removed from both cylinders, as the larvge had become full-fed. 
In each cylinder four green pupae were obtained, fixed to the brown stems or roof, 
or lying free on the floor. (One of these is figured in Plate 26, fig. 14.) These results 
surprised me very much, as I knew that there was a well-marked brown variety of 
the pupa not uncommon in this species. 
Of the remaining larvae two were placed (August 25) in a smaller glass cylinder 
(about 6 centimetres in internal diameter and 1 decimetre in height) covered with a 
single layer of green tissue-paper, and with a roof and floor of the same material, and 
with abundant food-plant. 
The last larva was placed under exactly similar conditions, but was blinded (renewed 
as above). 
This last larva and one of the former died, but the remaining larva pupated upon 
the green food-plant, and to my great surprise it produced a distinct broivn variety. 
(This pupa is represented in fig. 15, Plate 26.) At first sight these results appear to 
be extremely startling, especially when it is considered that the experiments in this 
paper were directed towards the investigation of an adjustable imitative resemblance 
which, when present, forms the highest culmination of this method of protection. But, 
in the first place, it was probable that the green cylinder was more shaded than the 
other, and hence the production of the brown form might be accounted for, the 
species being simply susceptible to comparative darkness or illumination, and pro¬ 
ducing its corresponding dark or light variety in obedience to the respective stimuli ; 
although the experiments showed that the ocelli have nothing to do with the suscep¬ 
tibility. Furthermore, the eight green pupge were produced from the largest and 
