760 THE AMERICAN NATURALIST. (VOL. XXXII. 
angle. Fitting in behind this arm and abutting against the 
postero-dorso-lateral, as shown by impressions on both, was the 
posterior lateral, a plate not hitherto identified as such. It may 
be that Newberry’s supposed “hyoid (?)”’ plate 1 occupied this 
space, but further comparisons are necessary to establish the 
truth of this inference. 
The side plates of the body being now fully accounted for, it 
may be asked why the name anterior lateral is not substituted 
instead of “clavicular.” The answer is that only one arm of this 
plate corresponds to the anterior lateral of Coccosteus, while 
the bifurcated arm represents something entirely different. 
Hence we must either go on calling the whole structure clavic- 
ular, or invent a new name for it; we prefer the former course, 
although technically the term is a misnomer. It will require a 
separate article to illustrate the relations of the bifurcated arm, 
and we will pass over this for the present, remarking only that 
the inner branch consists of a long thin blade which is probably 
homologous with the zz¢erlateral of Coccosteus, and the external 
curved branch has articulated to it distally a peculiar warped 
plate, supposed to have formed part of the modified branchi- 
ostegal apparatus. 
The effect of this orientation of the clavicular is to revolu- 
tionize previous notions as to the form of cross-section in 
Dinichthys and Titanichthys. Instead of being deep-bodied 
creatures, it is now plain that the more specialized species, 
with their flat dorsal and abdominal shields and excessively 
wide cranium, must have had almost ray-like proportions, and 
this depression of body was no doubt correlated with bottom- 
feeding. We observe also, which was not suspected before, 
that the plastron was not in contact with other dermal plates, 
and covered a relatively small portion of the abdomen. The 
flat portion of the suborbital probably had a continuous slope 
with the head shield, its inclination (and also that of the clavicu- 
lar) being more nearly horizontal than vertical. These rela- 
tions can best be shown in a side-view restoration, which we 
hope to present at a subsequent time. 
The only writer to attribute an opercular plate to Dinichthys 
1 Loc. cit. (1889), Pl. V, Fig. 3. 
