pact. BN 
ation, at the breaking-joint. At first I tried cutting off the legs 
quickly with sharp scissors. In a few cases the leg did not break off 
at once. Then the claws of the first leg were cut off so that the crab 
could not catch hold of the injured leg. Nevertheless in every case 
the leg was subsequently thrown off. I then tried the experiment on 
crabs that had just molted, and which were still soft. These too, sooner 
or later, threw off the legs at the breaking-joint. At last I hit upon 
a device that was successful. The crabs were taken from the snail- 
shells and put into sea water to which 5°/, alcohol had been added. 
In a few minutes the crabs were.so affected by the alcohol that the 
usual reflex did not take place when the legs were cut off. Moreover 
after the crabs had recovered from the affect of the alcohol the leg 
was still retained in several cases. I suppose that the nerve left ex- 
posed at the time of the operation had withdrawn somewhat from 
the surface and by the time the crab had recovered the stimulus was. 
no longer present. By operating on a large number of crabs a few 
were obtained that did not throw off the legs at the breaking -joint. 
There is another fact connected with the autotomy of this crab 
that is not, I believe, generally known. If the most distal segment is 
cut off the leg is not thrown off. This is also, in part, true for the 
segment next to the distal one; if cut off at its distal end the leg 
may or may not be thrown off; if the segment is cut off nearer to. 
its proximal end the leg is more apt to be thrown off. For the pur- 
pose of my experiment it was, therefore, necessary to cut off the leg 
at some point between the breaking-joint and the second segment from 
the end. 
All together 278 crabs were used in this experiment. In several 
cases large numbers died because the conditions were bad. Many 
crabs threw off the injured leg at the breaking-joint -after the oper- 
ation so that relatively few were left for observation. Occasionally 
a crab threw off the leg operated upon at the next moult at the 
breaking-joint, and a new leg regenerated from the base. Only five 
crabs regenerated from the cut-end and only in two did no re- 
generation take place. The difficulty of carrying out successfully the 
experiment is not due to the lack of power of regeneration from the 
cut-end, but is owing to the legs breaking off at the base sub- 
sequent to the operation. The result is, therefore, all the more sur- 
prising — or at least should be to those who hold to the hypothesis 
that I am combatting — for it can rarely if ever happen under na- 
tural circumstances that a limb can have an opportunity to regenerate 
outside of the breaking-joint (the last two segments excepted). 
