THE DEVONIAN ROCKS OF CANADA. 109 
tion of the dorsal not much less than the length of the head. In one specimen there are 
indications of what looks like a short break or separation between the anterior and poste- 
rior portions of the dorsal (or between the two dorsals, if there were two, which is 
scarcely probable) but this break may be owing to an accidental and abnormal fracture 
of the fin rays at this point, for in other specimens the two portions appear to be 
continuous. Anal and caudal fins both extending as far outward from the body as the 
posterior half of the dorsal does (assuming that there was but one dorsal), and separated 
at their bases by a very narrow interval. Anal fin narrow and elongated, ventrals acutely 
lobate and separated from the anal by a space considerably wider than that which inter- 
venes between the anal and caudal. Pectorals broader and longer than the ventrals, the 
former consisting of a long and acutely pointed scaly lobe, with the fin rays fringing it 
up to the base. 
Both the upper and under jaw are armed with smooth, conical and somewhat com- 
pressed teeth. The dental plates of the palate are each furnished with rows of erect, 
conical teeth, which seem to be arranged in the form of a rectangular patch or quadrature 
of a circle, very much as they are represented to be in Dipterus. As in the typical and 
previously only known species of the genus (the P. Andersoni of Huxley), the “ notochord 
was persistent throughout the entire length of the vertebral column, while the superior 
and inferior arches were well developed and thoroughly ossified.” The ribs are long, 
slender and well developed, and, like those of P. Andersoni, “stare through the integu- 
mentary scales of the fish” in the conspicuous way which suggested the generic name. 
The neural spines are elongated and slender, narrow in the middle and moderately 
expanded at each end. The interspinous bones, which support the fin rays of the dorsal 
surface are shaped very much like the neural spines to which they are adapted, but are 
a little shorter and straighter, their extremities being truncated at nearly a right angle 
to their longer axes. 
The specimens upon which the original description of this species was based were all 
collected by Mr. Foord in 1880. Although crushed flat laterally and considerably distorted, 
four of these specimens are very nearly perfect, but the pectoral and ventral fins are want- 
ing. The rest are mere fragments, but one of them shews the shape and position of one 
of the ventrals. The largest individual collected in that year is a little more than six 
inches long and three inches and a quarter in height or depth, while the smallest is about 
thirty-four lines long and ten high. The variation in the outline of different specimens, 
and in the proportions which their length bears to their height, is obviously due to the 
abnormal flattening to which they have been subjected. The smallest examples seem to 
be the least altered by this lateral pressure, and in these the length is much greater in 
proportion to the height than it is in the larger ones. 
In 1881 Mr. Foord collected a few additional specimens of this species, including one 
very large individual in which one of the pectorals and both of the ventral fins are pre- 
served in place. These shewed that the species attained to more than twice the size of 
the largest specimen previously collected, and exhibited very clearly also the character of 
the teeth in the jaws and palate. 
As compared with the P. Andersoni of Huxley, from the Old Red Sandstone of Dura 
Den in Fifeshire (which, as already stated, was the type and only known species of the 
genus, prior to the year 1880) the P. curtum, as its name was intended to imply, appears 
