VERTEBRATES. 125 



' sulcus are thin and sharp, and are usually more or less hroken. The smooth 

 and flattened surfaces of the base seem to indicate some kind of an articula- 

 tion, and it is perhaps possible that bony or cartilaginous guides were applied 

 on either side, between which the spines were firmly held during their eleva- 

 tion or depression. This would give an accuracy of movement and a firmness 

 scarcely inferior to that afforded by the gingylmus articulation of the dorsal 

 spines of the Siluroids. 



Figure 1, represents the side view of a spine broken at either end; 1 a, trans- 

 verse section; lb, an enlarged tubercle of the anterior border; 2, a nearly per- 

 fect base of another specimen. 



Formation and locality: Keokuk limestone. Keokuk, Iowa. 



Drepanacanthus? stellatus, N. and W. 



PL XII, Fig. 7. 



Spine small, strongly curved forward, with an elliptical sec- 

 tion above, oval below; convex margin rounded and ornamented 

 with three rows of relatively large, longitudinally remote, 

 compressed stellate tubercles; concave margin rounded and set 

 with a single row of large, conical, pointed, stellate tubercles, 

 largest toward the base of the spine; lateral surfaces regularly 

 arched, with numerous longitudinal rows of small, conical or 

 laterally compressed tubercles, most distinct on the* anterior 

 border, separated longitudinally by intervals of twice their 

 diameters, laterally closely approximate, forming oblique, trans- 

 verse lines. In the middle of the spine there are about fifteen 

 longitudinal rows of small tubercles on either side, more below, 

 fewer above. The large tubercles of the convex (posterior?) 

 border are separated longitudinally by spaces of twice their 

 longest diameters. Transversely they are in apposition, form- 

 ing short transverse lines of three tubercles each. 



We have included this spine in Drepanacanthus with some hesitation, as the 

 imperfection of the only specimen in the collection is such as to leave several 

 important points in its structure in doubt If, as we have supposed, the curve 

 of the summit was forward, as in L. gcmmatus, it would naturally fall into the 

 same genus with that spine; for although the tubercles of the concave border 



