30 
arrangement based upon scale- cliaracters. He points out 
the many fatal anomalies which result from the definitions of 
orders according to Agassiz, and shatters the delusive hope 
of establishing a natural history of value upon such superficial 
distinctions. The essential peculiarities of the Ganoids are 
shewn to lie, not in scales, fins, or tails, but in the heart, 
brain, and intestine. 
The anatomical research of this paper gives it a perma- 
nent value which the labours of no mere systematist can 
confer. The structure of Pohjpterus and Lepidosfeus in par- 
ticular is given with a fullness and completeness which 
more recent investigation has hardly modified. In the re- 
construction of the order certain of the Siluroids, Scleroderms, 
Grymnodonts, and Lophobranchs admitted by Agassiz as 
armour-plated fishes, are removed entirely from the order 
Ganoidei. The remaining recent genera are enumerated by 
Miiller thus: — Polyjoterus^ Lepidosteiis, Accij^enser, ScapM- 
rhynchuSf and Bpatularia {Poly don). 
Since 1844 the list of genera has been modified as follows : 
— The Ganoid structure of Ainia, discovered by Yogt, is an- 
nounced in a postscript to Muller's paper. The Amia (Bow- 
fin or Mud-fish of the United States) is a Ganoid, disguised 
as a herring or sprat. Ganoid in its heart, optic nerves, and 
intestines, it has soft cycloid scales, a fact which in itself 
suffices to vitiate the systematic value of scale characters, 
CalamoichthySf first described in 1865, and CeratoduSy added 
to the list of living genera in 1870, were unknown to Muller. 
Lepidosiren, ranked by Agassiz as a Ganoid in his later mono- 
graph, but excluded by Muller, it is now proposed to re- 
admit. The conspicuous difierences of their structure from 
that of the typical Ganoids have been bridged over by the 
discovery of Cemtodus, as will shortly appear. In other 
respects Muller's restricted order is co-extensive with that of 
the ichthyologist of 1872, and the ordinal definition (with 
