554 THE AMERICAN NATURALIST. (VOL: XXXII. 
an equilateral triangle, but the side corresponding to the poste- 
rior face in other forms is deeply excavated, the depth of the 
wedge-shaped cavity amounting to almost half the height of the 
triangle. In the view given of this side (Fig. 50, Z), an attempt 
is made to show the change of slope and median groove at 
the bottom of the cavity. The surface of the latter is smooth 
throughout, and the appearances are unmistakable that it either 
contained soft parts or was attached to them during life. Such 
could hardly have been the case, however, if the spine stood 
erect and free from the body of the animal, but would neces- 
sarily happen were we to suppose it imbedded lengthwise in 
the flesh. That it actually was so imbedded appears the more 
probable when we consider the external aspect. Parallel mark- 
ings are seen along the border on the lateral faces which 
apparently indicate the limits of integumentary covering; above 
these markings the sharp angle of the wedge protruded free as 
a cutwater, and probably served also as an offensive weapon; 
below them the bone sank beneath the skin, and was firmly 
secured by muscles attached to the channeled face. The 
latter, in the position suggested, would be zzferior instead of 
posterior; the smaller, pointed end would be posterior instead 
of distal; and the larger, heavier end would be anterior instead 
of proximal or inserted. 
Not only are the mechanical difficulties much lessened, of 
supporting free from the body an organ of such size and weight 
as this, according to the interpretation just outlined, but its 
plausibility is strengthened by the analogy of Edestus, with 
which it agrees in having no medullary canal. There is no 
doubt that the simpler types of Edestus (E. heinrichsi) were 
principally imbedded in the integument, so that only the row 
of denticles protruded. Bashford Dean,! in a very luminous 
paper on this genus, concludes as follows regarding the origin 
of the structure: ‘In the present case the evidence may be 
accepted as conclusive that a spine-like organ had its origin as 
a metameral structure whose basal portion lay within the integu- 
ment, and traversed longitudinally a number (seven at least) of 
1 Trans. N. Y. Acad. Sci. vol. xvi, p. 68, 1897. Cf. also Fishes, Living and 
Fossil, pp. 28-30, 1896. 
