METAMORPHOSIS OF ACTINOTROCHA. 218 
are extremely different in appearance and in structure. Each 
is very perfectly adapted to its own special set of conditions ; 
and since the two sets of conditions are very unlike, the cor- 
responding structural adaptations are also very unlike. Now 
the animal must somehow exchange the one set of characters 
for the other; and of course the most advantageous method 
of exchange will be preserved. If the larva were converted 
into the adult by a process of continuous growth, gradually 
losing its larval peculiarities while acquiring the adult 
feature, evidently an intermediate or transition period would 
result, during which the animal would be imperfectly 
adapted to either larval or adult life, and yet would require 
all its energies to keep up the structural changes going on. 
It would either be at a great disadvantage during this period 
or would have to acquire still another set of compensating 
adaptations answering to the requirements of the transition 
stage. Examples of such adaptations are common, e.g. 
among many insects where the changes undergone are so 
profound, apparently, as to render transition stages inevi- 
table. 
For these reasons, it seems manifest that it would be 
advantageous for the animal to pass rapidly and directly 
from the larval structure to the adult, thus escaping as far 
as possible the disadvantages entailed by a state of transition 
or the expenditure of energy demanded by the acquisition 
of secondary adaptations. The metamorphosis of Actino- 
trocha secures to it this advantage, and it seems clear that 
this is why the process of sudden transformation has been 
acquired. 
Passing on now to consider how the process may have 
been acquired, the bearing of the speculations concerning 
the way in which the anus originally left its primitive ter- 
minal position becomes apparent. If my theory of this 
process is correct, it is evident that in the adult worm the 
dorsal region is nearly obliterated, being represented only 
by the short interval between the mouth and the anus; 
while the ventral region includes nearly the whole length 
of what is apparently the dorsal surface as well as the whole 
“ventral”? surface. In the larva, however, the dorsal and 
ventral region are of nearly equal extent; so that in assum- 
ing the adult condition the ventral region must undergo 
extensive increase through some process of growth. From 
what has been said above, it is evident that advantage would 
be derived if this growth could take place in such a way as 
not to impair the adaptation of the larva to its conditions of 
life. A simple ventral outgrowth of the body wall, carrying 
