762 THE AMERICAN NATURALIST.  [Vou. XXXII. 
original head of T. agassiziz, and enough of the sutures are 
traceable to give a pretty fair idea of the arrangement of 
cranial plates. The boundaries of the latter are represented 
in the accompanying diagram (Fig. 4) by continuous lines as 
far as they can be made out with certainty, and are dotted in 
where more or less obscure. Sensory canals are indicated by 
the usual convention of double dotted lines. The posterior 
and lateral margins of the skull are entire; the part broken 
away includes a portion of the preorbitals and pineal and the 
whole of the rostral (or “ethmoid”) plate. It is very evi- 
dent, however, that the head was more elongated from side 
to side than in an antero-posterior direction, which is opposite 
to the usual rule. It is also almost perfectly flat, instead of 
being transversely arched. 
In consequence of the huge size of the head and thinness of 
the plates, rigidity could only be attained by a nearly com- 
plete fusion of the cranial elements, and this rendered the 
artifice of dovetailing unnecessary. Nevertheless a vestige of 
the usual interlocking condition remains in the anterior bound- 
ary of the central element, where there are a few moderate- 
sized undulations. Presumably the centrals met each other in 
the middle along a wavy line, as mechanical principles would 
seem to require, but the suture itself is now almost wholly 
obliterated. There is likewise no sort of indication that the 
space homologous with the central was divided into two com- 
ponents. The pineal plate seems to have been of elliptical 
outline and longest in a transverse direction. It is extremely 
attenuated at its lateral edges, where it has been somewhat 
broken away. Here and along the antero-external margins of 
the central were the thinnest regions of the cranium, the bone 
being considerably less than .5 cm. through. The pineal fora- 
men is enclosed in an elliptical capsule of very dense tissue, 
and opens on the dorsal surface in a circular orifice. An 
interesting fragment in the collection of Dr. William Clark, of 
Berea, Ohio, shows that this opening was covered by a thin 
opercular plate lying loose on the upper surface and undoubt- 
edly movable in life. In another specimen the foramen is seen 
to be double, the two being separated a slight distance sidewise. 
