206 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



branches. The Burlington species of Physonemus and Stethacanthus display 

 a marked increase in size, but they are feebly ornamented, and remain so 

 throughout the stage represented by the Keokuk limestone. Stethacanthus 

 seems to have attained its maximum size in the Keokuk Group, as Physo- 

 nemus did in the Burlington, and a considerable falling off in this respect is 

 true of both genera in the St. Louis division. The spines of Stethacanthus 

 remain unornamented from the time of their first appearance in the Berea 

 grit of Ohio until their extinction near the close of the Subcarboniferous, 

 but those of Physonemus and Erismacanthus increase in complexity of or- 

 namentation throughout the Mississippian series, ultimately displaying remark- 

 able elaboration. An inspection of the forms illustrated in Plate XXII. of the 

 sixth volume of the Illinois Palaeontology, and of the spines figured in the 

 present contribution, will convince any one as to the correctness of these gener- 

 alizations. 



The spines in the typical species and in others resembling it are much later- 

 ally compressed, strongly arched or hook-shaped, with a broad base of inser- 

 tion ; the sides of the exserted portion are more or less ornamented with 

 tuberculated longitudinal ridges, and small denticles are present upon the 

 concave (posterior) border. This description applies to P. arcuatus M'Coy 

 (the type species), P. attenuatus Davis, and P. hamatus (Agassiz), from the 

 Carboniferous Limestone of Great Britain ; and to the American forms de- 

 scribed as P. stellatus Newberry, and Drepanacardhus reversus St. John and 

 Worthen. Another group of spines which may be referred provisionally to 

 the same genus is typified by such forms as the so-called Drepanacanthus 

 gemmatus Newb. and Worth., D. anceps Newb. and Worth., Xijstracantlms 

 acinaciformis St. J. and Worth., Physonemus gigas Newb. and Worth., and the 

 defences Lheoretically associated with the teeth of Polyrhizodus rossicus by A. 

 Inostranzew ^ and 0. Jaekel.^ It is characteristic of the latter group of 

 spines that they are forwardly curved, instead of backwardly, as in most 

 Ichthyodorulites, a circumstance which appeared so anomalous to Newberry 

 and Worthen as to warrant a generic separation from Physonemus. Transi- 

 tional stages, however, showing the reversal of curvature from a posterior to 

 an anterior direction, are to be observed in various species of Stethacanthus 

 and Oracanthus, and for the present it seems best to extend the definition of 

 Physonemus so as to include both groups. The two rod-like species from the 

 Kinderhook limestone immediately to be described diff"er from all others in 

 their more slender form and absence of ornamentation. They are undoubtedly 

 to be interjireted as head-spines, a determination which is applicable to nearly 

 all species of this genus. 



1 Travaux See. Nat. St. Petersb., Vol. XIX., 1888, pp. 1-18. 



2 Zeitschr. deutsch. geol. Ges., Vol. LI., 1899, p. 281, Fig. 5. 



