2 
summer I made at Woods Holl a new series of experiments, which, as I 
to show, give a crucial test of the current dogma. 
It is known that there exists at the base of the walking legs of 
the hermit-crab an arrangement by means of which the leg, if injured, 
may be thrown off. The breaking-joint lies between the fifth and sixth 
segments counting the distal segment as the first. These segments are 
not free to move on each other, and do not, in one sense, represent a 
weak link in the chain of segments; for if the crab be killed the leg 
can not be more easily broken off at this joint than elsewhere. As 
FREDERICQ has shown the living crab breaks off its leg by some sort 
of reflex action. The exposed surface left at the breaking-joint is 
covered by a thin cuticular membrane that is pierced at its center by 
a small hole through which the nerve and blood vessels of the leg pass. 
If the leg is injured outside of the breaking-joint (except near the 
tip) the leg is promptly thrown off at the breaking-joint. Sometimes, 
however, if the leg is cut off quickly with sharp scissors it may not 
be immediatly thrown off, but in such cases the crab catches hold of 
the injured leg with its claws and by holding the broken stump of 
the leg the resistence is sufficient to allow the usual mechanism to 
come into play, and the leg is thrown off. 
The special problems that I have examined are these: if the leg 
is cut off proximal to the breaking-joint, will a new leg develop; if the 
leg is cut off distal to the breaking-joint, and can be prevented from 
being thrown off at the joint, will regeneration take place from the 
exposed end? 
The Legs (I, II, II) cut off proximal to the 
Breaking-Joint. 
With a pair of small scissors it is possible to cut .off the leg in- 
side of the breaking-joint. The leg was first injured so that it was thrown 
off at the breaking-joint, then the part left was cut off close to the 
body—usually in the middle of the proximal segment. The part cut off 
was examined to see if it always contained the outer part of the prox- 
imal segment and the ring-like following segment. In all 208 crabs 
were operated upon in this way. Some of the crabs died as a result 
of the operation; others died because at first the water in which they 
were kept was impure, but in nearly all those that survived regeneration 
of the missing limb took place. 
In the most successful experiment (Aug. 24—Sept. 26) in which 
25 crabs were used, the smaller first leg being cut off inside of the 
breaking-joint, regeneration took place in 15 individuals; it had not 
