354 Mr. Osler on burrowing 
becomes an oval, when the anterior points of the valves are 
approximated. The ventral margins have also a consider- 
able convexity. At a fourth of the length of the valves from 
their anterior points, their edges are rolled in to form an 
umbo, which is covered with a reflected semi-pearlacious fold ; 
and at the inner part of this is a round, prominent, cartila- 
ginous spot, the point of union between the valves. In P. 
Candida, a short process is given off from beneath this spot 
in the left valve, (fig. 4, a) and a thin, flat ligament from a 
corresponding tubercle in the right, which, passing obliquely 
backward, meet at the middle of their length, and are thence 
united to their extremities. The connection thus obtained 
is so weak, that it is extremely difficult to remove the fish 
without separating the valves : it is, indeed, only strong 
enough to make the hinge the constant centre of motion. 
A long, thin, curved process is given off from beneath the 
umbo, (fig. 4, b). Its convexity corresponds with that of 
the valve, and its direction is a little obliquely backward. I 
believe this process to be characteristic of an animal which 
bores mechanically by employing its shell as a rasp. The 
anterior and lower portion of the shell is thicker than any 
other part, and its surface is armed with much stronger 
spines ; and it is this portion which the Pholas employs in 
boring. 
The Pholas is quite destitute of the elastic ligament which 
connects and expands the shell in other Bivalves. Its con- 
nections are chiefly muscular, and its motions are effected 
exclusively by the action of muscles. It is evident that, by 
this arrangement, the greatest possible freedom of motion is 
secured. 
