MITCHELL — ON THE RESTORATION OF PTERASPIS. 405 
servation, and their exact measurements have been followed in the 
figure. The first specimen, used for this diagram, exhibits very 
beautifully the form of the shield with the terminal horns, and the 
distinct eye-sockets. The eyes are placed on the margin of the shield, 
and their impression is also seen on a cast in the stone of this spe- 
cimen. The second specimen 
is the prolonged central ter- 
mination of the shield, which 
has been broken off at the 
ridge which terminates on 
either side in the horns. The 
third specimen shows the junc- 
tion of this central prolonga- 
tion with the shield. All the 
three specimens have a high 
central ridge, and still retain 
something of the graceful out- 
line of the living form. As 
preserved in the stone these 
specimens show only the na- 
creous layer, the other com- 
ponent layers of the bone of 
the Pteraspis having perished 
in their case. 
Figure 2. Two specimens 
have been employed for the 
construction of this diagram, 
the one a piece of solid bone, 
and the other a cast in the 
stone of the under surface of 
the dorsal shield. They are 
apparently the relics of indi- 
viduals of the same species and 
even size ; and we have again 
followed their exact measure- 
ments in the figure. The spe- 
cimen of solid bone exhibits 
the external aspect and general 
Fig. 2. 
outline of the shield, and is especially valuable as showing the form 
and position of the horns of the shield. It will be observed that 
these are not the prolonged cusps of the Cephalaspis, or at least 
of the Cephalaspis Lyellii. The cast shows the marks of the eye- 
sockets, and exhibits traces of the central ridge of the shield. The 
ridge, however, is a low one in comparison of that in Figure 1 ; 
and we have hitherto failed to discover the central prolongation or 
termination of this larger or adult form. 
Figure 3. This must be understood to be very conjectural; and 
although three of the plates are copied exactly as they occur in the 
stone, their arrangement aiid their function must be considered so far 
