246 
Proceedings of the Royal Society of Edinburgh. [Sess. 
away, having been divided, as before, at the breaking-plane in the basi- 
ischium, but retained in its position relatively to the animal on account of 
the tight packing and folding of legs within the mouth of the shell. 
On account of the small size of the crab, and the covering afforded to 
the basal limb-segments by the shell, I was unable to study the process of 
autotomy by graphic methods under normal conditions. It was found, 
however, that when the shell was broken away over the tail, and the 
nerve-segments of the particular limbs operated on separated from those 
nerve-segments above and below in the gangliated chain, autotomy took 
place. Under normal conditions, therefore, the reflex of autotomy in the 
hermit crab is unisegfinental. This is a distinct advance on the condition 
found in the Macrura, for there it was found that the reflex involved the 
nerve-segment controlling the damaged limb, and the ganglia and arcs of 
the abdominal region also. 
It is practically an unknown thing to find a hermit crab without a 
shell. The animal, during its growth, has to change shells periodically, 
but it takes very great care to find a suitable new home before it leaves 
the old one. It places the new one close up against the mouth of the one 
it is about to discard, and those who have seen the changing report that 
it only occupies a portion of a minute. Autotomy is therefore occurring 
under abnormal conditions when it happens in a crab removed from its 
shell ; but it is a most surprising thing to find to what extent the reaction 
of the animal to injury varies. Instead of extending its limb or contracting 
its tail, the hermit crab promptly plucks oft" the stump of a damaged leg 
with its chelipeds. The reaction is deliberate and purposeful. Division 
always takes place at the breaking-plane, and the profuse hsemorrhage is 
at once stopped by closure of the valve which is formed by flaps of the 
membrane obturatrice.” But the influence of the cerebral ganglia does 
not in the least affect the reaction. Division of the nerve-cord below the 
level of injury is likewise without effect. 
It was stated that no marked extension of the leg can be noticed ; 
but if the extensor muscle of the second segment be connected up to a 
heart lever, it is found to contract vigorously when the limb is cut, 
burned, or crushed, or when the cut end is stimulated by a faradic current. 
The latent period is about one-tenth of a second, and the period of re- 
laxation is about fifty times as long as the period of contraction. This 
contraction will not take place if the nerve trunk of the segment be divided 
as it leaves the central nerve mass. It is, therefore, a reflex. Since no 
marked extension of the limb can be observed, there must be a flexor 
antagonism to this extensor contraction, but no details of such could be 
