REGENERATION OF LOST PARTS IN ANIMALS. 41 
regenerate the dorsal surface, the converse does not occur. 
It might be said that the ventral surface, with its tube- 
feet, water-vessel, nerve-strand, etc., is much the more 
complex; but, on the other hand, it seems justifiable to point 
out that the dorsal surface is, in natural conditions, much 
more open to attack and injury than the ventral surface 
adhering to the rock. Again, perhaps, the exception proves 
the rule. 
Another apparent difficulty which turns out to be a 
corroboration is expounded by Bordage. It is perhaps a 
little difficult to follow without diagrams, but it is worth a 
little trouble. The lower or tarsal joints of the legs of 
locusts and the like are not difficult to dislocate, and in the 
two front pairs they are readily regenerated. But if the 
leg be tried higher up, it requires great force to break it 
between parts 1 and 2 (coxa and trochanter), or still more 
between parts 2 and 3 (trochanter and femur). The injury 
is often fatal, but if the insect survives and is young, 
regeneration may be effected, especially if the separation 
was between parts 2 and 3, where it is with most difficulty 
brought about. Now, after making all allowances for the 
various ways in which locusts may be pulled about by one 
another, or by birds and other enemies, it seems difficult to 
see how in natural conditions sufficient force would be 
exerted to break the leg, and still more difficult to under- 
stand why regeneration should most frequently follow when 
the breakage occurs at the most difficult place. Yet the 
difficulty is premature, for observation of the frequent 
moultings shows that when the locust is struggling out of 
its old clothes, breakage at the joints is very apt to occur, 
particularly at the trochanter-femur articulation, which 
afterwards becomes so strong. The difficulty disappears and 
becomes an argument in favour of the adaptive nature of 
regenerative phenomena. Similarly, Bordage has shown in 
regard to Phasmids, where the assaults of birds and lizards 
seemed insufficient reason for the prevalent habit of breaking 
off a leg at a particular line and regrowing it thence, that 
the breakages during emergence from the egg or from the 
cast husks have probably furnished sufficient reason for the 
evolution of the restorative provision. Of a hundred 
