528 MULLER ON THE STRUCTURE OF THE GANOIDS, 
row of fulcra on the upper and under margin of the caudal fin 
of Tetragonolepis and Ptycholepis as far as the extremity. They 
also appear to be simple in the thoracic fins of this genus, ac- 
cording to Agassiz’s plate of Tetragonolepis confluens (Ag. ii. 
tab. 23a, fig. 1). Pholidophorus also appears to belong to this 
place, from the fulcra on the upper and lower margin of its tail-fin. 
In other genera of Ganoids the anterior margins of the fins are 
covered with a double row of fulcra, just as we see among living 
Ganoids in Lepisosteus; exactly the same occurs in all the fins 
of Lepidotus and Caturus. That the same applies to the thoracic 
fins of Lepidotus, is shown in the figure of Lepidotus Mantellii, 
Ag. (See Agassiz, ii. tab. 30c.) I find the double rows in 
this and in all the other fins, as also on both margins of the 
caudal fins. In a large species of Catwrus, from the Lias of Boll, 
which is probably the Caturus Meyeri, V. Miinst., I find several 
strong undivided fulcra at the commencement of the caudal fin ; 
but these immediately pass into double rows of fulcra, which 
cover the entire length of the anterior margin. I also find these 
double rows of fulcra on the fins of Pachycormus macropterus, 
Ag., in those parts where fulcra are present, 7. e. on the dorsal 
and anal fins. Semionotus has also double rows of fulcra (pec- 
toral fins). These differences indicate essential variations ; it is 
impossible to find a more remarkable distinction than the caudal 
fin of Ptycholepis and Tetragonolepis, with simple rows of spinous 
fulcra, and of Lepidotus and Lepisosteus, with double rows. In 
Pachycormus the double rows occur with an unossified condition 
of the centre of the spinal column; but in Lepisosteus, and 
apparently also in Lepidotus, with an osseous spinal column, 
According to Agassiz (Ad. tab. 29c, fig. 12), the genus has a 
perfectly ossified spinal column, and thus makes an exception 
to the other Lepidoidei, in which, according to the same author 
(l. c. 182), as far as he is acquainted with the remains of the 
skeleton, the vertebral column is wanting. 
Although the heterocercous Ganoids are much more nume- 
rous in the older formations, still all the fish found in them are 
not heterocercous. It is a very remarkable circumstance, that 
the forms belonging to the families Lepidoidei and Sauroidei, 
Ag., which occur prior to the Jura formation, are heterocercous, 
as Agassiz has pointed out ; but this is rather a consequence of the 
system than of natural relations ; the result is at once destroyed, 
when we take into consideration Celacanthus and Undina, which 
