RELATIONSHIP BETWEEN LARVAL AND IMAGINAL LEGS OF LEPIDOPTERA, 143 
So far back as 1894, however, J. Gonin published a most valuable 
paper on the metamorphosis of lepidoptera in the Bulletin de la Société 
Vaudoise des Sciences Naturelles, in which, amongst a most excellent 
series of observations on the development of the appendages of Pieris 
brassicae, for nearly all of which I have nothing but praise and agree- 
ment, he adopts with regard to the legs the second answer, v7z., that. 
the imaginal legs entirely originate from imaginal discs within the 
body of the larva. Or he may, indeed, be understood to say that the 
imaginal tarsus arises from the larval leg, and some of his expressions 
imply that he regards the three joints of the larval leg, not as being 
respectively femur, tibia, and tarsus, but as being actually three 
joints of a tarsus, the representation of tibia and upwards being within 
the larval body. 
The question is thus expressed by Gonin (Bull. Soc. Vaud., vol. 
xxx., 94), that Reaumur has been misrepresented, and that his actual 
words are, that he cut off ‘‘more than the half of three of the true 
legs,” and found that the chrysalid had ‘‘the three legs of the side 
shorter than the corresponding limbs of the other side,” and that a 
larva experimented on at a younger stage showed a fresh three limbs 
in the pupa, but “atrophied,” that is to say not entirely absent, and 
he criticises Ktinckel for saying that it is clear that ‘‘ Reaumur, having 
completely cut off one of the true legs in some caterpillars, proved that 
the butterfly that emerged was without the corresponding appendage,”’ 
and says that Newport denied this disappearance of the legs, and 
regarded the limb as partially regenerated. He then goes on to 
describe the state of matters at the date of the change to pupa, when 
the greater part of the lee has so far assumed its imaginal character, 
and increased in size, as to have left the larval leg, and to be pressed 
together at its base within the larval skin, leaving only a portion in 
the larval leg roughly corresponding to the tarsus, and, though he 
seems to have a good grip of the processes by which the imaginal lee 
arises, he appears in a fashion curiously illogical, considering his 
actual knowledge, to believe that the condition at this date, 7.e., the 
moment preceding the moult to pupa, represents the true relations of 
the larval to the pupal (or imaginal) leg, viz., that the imaginal leg 
arises from larval structures situated where he finds it at this late 
period, when it has really by growth left its confined quarters in the 
larval leg, and that the larval leg corresponds to the extremity (say 
tarsus) only of the imaginal limb, and that the rest exists in the larva 
merely as an imaginal disc not within the larval leg, but in the body 
of the larva at its base. 
Newport’s experiments, fully recorded in the Philosophical Trans- 
actions for 1844, relate to Aylais urticae, and, properly interpreted, 
seem to render Gonin’s position untenable, whilst the results correspond 
entirely with those of the experiments I made last year; he brings out 
a point with which I did not meet frequently enough to note it 
definitely, and that is, that removal of a portion only of the leg results 
in the tarsus being reproduced in an incomplete state, that is with less 
than five joints. His observations on the reproduction of spines and 
spurs do not quite accord with mine, but this also seems to have 
occurred in the case of partial removal. Newport’s experiments were 
made with a view to learn whether regeneration of amputated parts 
occurred in lepidoptera ; mine were chiefly with a view to test Gonin’s 
