462 Mr Brindley, Regeneration in Sarnia ailanthus. 
The results obtained under ( B ), ( C) and (D) above differ from 
those of Newport and Chapman in two respects, viz. the terminal 
claw apparatus was as a rule not present in the regenerated limb, as 
it was characteristically in their cases, and in Sarnia there was no 
case in which the tarsus was regenerated with the normal number 
of joints, while Chapman has figured at least three cases of a 
5-jointed tarsus in Liparis. It is quite possible that generic 
differences exist in these respects. The experiments on Samia 
tend to suggest that the earlier the instar injured the more the 
imaginal limb approaches the normal in form and size. As 
regards the comparative results of injuring a leg to a greater or 
lesser extent at the same larval stage, the results obtained under 
( G ) and (!)), which it was hoped would throw light on the point, 
were attended with so much variability, and were, owing to the 
numerous deaths in pupa, so few in total number, that all that 
can be said is that there is at least a suggestion that an injury 
at the onset of pupation is, if confined to the middle joint of the 
leg, followed by the production of a rather better formed imaginal 
leg than if it includes removal by amputation through the basal 
joint. So far as it goes, there is here some evidence in favour 
of Gonin’s conclusion 1 , based on anatomical examination, that just 
before pupation only the extremity of the developing pupal leg 
projects into that of the larva, and that amputation of the latter 
at its base is therefore likely to injure the pupal leg more exten- 
sively than section through the middle joint. Chapman’s observ- 
ations on Liparis were undertaken in part with the object of 
examining the evidence for Gonin’s view, and he arrives at a con- 
clusion unfavourable to this author, but I am unable to agree with 
the interpretation he places on his results with Liparis in this 
respect, though, as stated above, my own experiments on Samia 
are too few to do more than suggest that Gonin’s view is the 
correct one. 
1 Bull. d. 1. Soc. Vaudoise d. Sci. Natui'., 1894, s<5r. 3, xxx., p. 122. 
