THE DEVONIAN EOCKS OP CANADA. 95 



The few detached plates yet found are rarely perfect, though the sculpture of their outer 

 surface is generally well presented. 



In some respects, the Campbellton Coccosteus very closely resembles the C. cuspidatus 

 of Agassiz, but in others there are such marked differences between the two forms that it 

 is thought most prudent, for the present, to distinguish the Canadian species by a local 

 name. No detailed description of C. cuspidatus has beeu ever published and the illus- 

 trations that give the best idea of its character are the figs, on Plate III of the " Old Red 

 Sandstone." Assuming that these figures are essentially correct, the shape of the post- 

 dorsomedian plate of the Campbellton Ccoccosteus (which Agassiz, who calls it the dorsal 

 plate, regards as offering one of the best specific characters) and that of the diamond 

 shaped ventro-median are almost exactly similar to those of C. cuspidatus. But on the 

 other hand, in many of the plates of C. Acadicus, and especially in some which have not 

 been separately described, on account of the uncertainty of their homologies, bvit which 

 are supposed to be isolated dorso-median plates of exceptionally large individuals, the 

 tubercles are arranged in very distinct concentric lines, with continuous and compara- 

 tively broad grooves or spaces between them — an arrangement not indicated at all, or at 

 most very obscurely, in the figu.res of C. cuspidatus. Again, the su.perficial grooves on the 

 cranial shield of C. Acadicus are much more like those of C. decipiens as represented in a 

 woodcut in the "Foot-Prints of the Creator," (third edition, fig. 11) than they are like 

 those in the figure of C. cuspidatus in the " Old Red Sandstone." In the C. Acadicus the 

 most conspicuous of these grooves are constantly those which run from a to e on the ac- 

 companying diagram, and from the centra of each of these lines to the lateral notches 

 at b.b. Making allowance for distortion, x^reciscly similar grooves are to be seen in 

 Miller's woodcut of the " cranial buckler " of C. decipiens, but they are entirely absent 

 in his figure of the cranial shield of C. cuspidatus. Further, in the Campbellton Coccosteus 

 other superficial grooves run from e.e. and d.d. to //. in such a way as to inclose a 

 triangular space on either side, with a space between their inverted apices at //. This 

 again, is just the arrangement in the " cranial buckler " of C. decipiens, whereas in C. 

 cuspidatus the apices of the two triangles arc not separated by a space but connected by 

 a curved, transverse groove. It would seem, therefore, that the C. Acadicus may be dis- 

 tinguished from C. decipiens by the different shape of the j)ost-dorsomedian plate, from 

 C. cuspidatus by the different arrangement of the grooves on the outer surface of its 

 cranial shield, and from both by the peculiar sculpture of its bony plates. 



Ctenacanthus latispinosus, Wiiiteaves. 

 (Plate X, figs. 3, 3 a, b.) 



Ctenacanthus latispinosus, Whiteaves, 1881. Can. Nat. and Quart. Jouru. Sc, N. S., Vol. 



X., p. 99. (Compare C. ornatus, Agassiz. Rech. sur les Poiss. foss., Vol. Ill, p. 12, 



tab. 2, fig. 1.) 



Fin spines small, as compared with those of most of the other species of the genus, 



compressed laterally ; either elongated, slightly curved and tapering rapidly from a 



rather broad base to an obtuse point, or comparatively short, straight and triangular. 



Posterior margin somewhat concave and bearing on its upper portion certainly one row 



