618 J. W. DAVIS ON REMAINS OF FOSSIL FISHES 
0-7 inch at the median portion, from which the spine tapers 
rapidly anteriorly to an acutely pointed apex; in the opposite direc- 
tion the basal portion decreases in diameter with a somewhat 
pointed termination. This specimen does not indicate what length 
of the basal portion was open; but a second one, which is in great 
part fractured and exhibits a longitudinal section of the spine, 
appears to show that the basal portion imbedded in the body of the 
fish had an open cavity for about an inch from the basal end on the 
outer margin. The walls of the spine were thin, and in the 
specimens are crushed; they have a large internal cavity, which 
extends to the apex of the spine. The surface exhibits traces of 
strie; and a number of tubercles, apparently without definite 
arrangement, are scattered over the upper part of the spine. Ich- 
thyolites, apparently of the same species as this one, obtained 
from the Mountain Limestone of Bristol, were regarded by Prof. 
Agassiz as spines of the genus Onchus. The spines, however, are 
so divergent from the characters of that genus, that there can be no 
hesitation about removing the species and placing it with others 
in the genus Physonemus. 
This species differs considerably from the type species of the 
genus, viz. Physonemus subteres, Agass., and P. arcuatus, McCoy. 
The former consists only of a fragment of a spine from the lime- 
stone of Armagh, which is a very doubtful example of the genus. 
The latter, also from the Armagh limestone, was described by Prof. 
McCoy (Brit. Paleeoz. Foss. p. 638, pl. 31. fig. 29). It is much 
larger, less arched, regularly tuberculated on the surface, and has a 
double row of denticles on the posterior concave margin of the spine, 
each particular distinguishing it from the species now described. 
Messrs. St. John and Worthen describe several species of this 
genus from the Kinderhook limestone (Paleontology of Illinois, 
vol. vi. p. 451, pl. xvi. figs. 1, 2). The specimens are for the 
most part fragmentary, and have been much abraded; they are 
consequently somewhat difficult to determine. Physonemus proclivus, 
St. J. and W., approaches nearer than any other form to the one 
here described. It is not, however, nearly so much curved, and, in 
proportion to its size, appears to have been thicker and more robust. 
The anterior apex is shorter and less acuminate. The same remarks 
apply with greater force to Physonemus parvulus, St. J. & W., from 
the fish-bed horizon of the Keokuk limestone (op. ez. p. 453, 
pl. xvii. figs. 11, 12). 
Genus GompnacantuHts, Davis, g. n. 
Spine broad at the base, rapidly contracting in width towards 
the apex, which is pointed. Transverse section more or less circular. 
Internal cavity probably terminal, este  iuaecers sulcus. Surface 
uniformly striated. * 
GompHacantuus Acurus, Davis, sp.n. (Plate XXVI. fig. 9.) 
Spine small, 1°25 inch in length, greatest width at base °25 
of an inch; imperfect, tapers rapidly upwards to the point; upper 
