Considered by Dr, Christian Liltken. 



431 



would be at all natural. That no sharp line of demarcation divides 

 the palfeocercal and neocercal types is well known. Sub-homo- 

 cercal LepidosteidtB appeared as early as the ''Dyas," and true 

 heterocercal genera continued as far down as the Lias. Like the 

 Ganoids generally, so also this sub-division cannot be defined by 

 any single, precise, or absolute character ; the " fulcra " are also 

 found in ancient Teleosteans and Lepidopleuridce, the rhomboidal 

 articulated scales in many Grossopteri, etc. ; nor are the fulcra 

 discovered in all Lepidosteidce} This division can only be defined 

 in a negative manner, as comprising those rhombiferous Ganoids 

 which have neither the distinctive marks of the Lepidopleurid^d nor 

 those of the Grossopteri, but which forms nevertheless, as far as 

 I am able to judge, a very natural assemblage of genera. 



B. Lejndopleurida or the Pycnodonts, easily recognisable by their 

 dermal ribs, and by the peculiar manner in which their scales (when 

 present) are interlocked and attached to those ribs ("lepidopleura"). 

 Each of the three tribes composing this sub-division has its peculiar 

 geological distribution : 



c. Fycnodontes 

 a. Fleurosomi b. Pleurolepides (non-fulcrati, homocerci.) 



(fulcrati, heterocerci.) 

 Carboniferous, Permian. 



(falcrati, homocerci.) Mesozoic Neozoic 



(Liassic, (Cretaceous, 



Liassic. Oolitic, Eocene). 

 Cretaceous).^ 



The systematic unity of these tribes has been attacked by Heckel 

 and Wagner (two of the chief authors on this group), but it is put 

 beyond doubt by the excellent writings of Egerton and Young. 



C. To the great mind of Professor Huxley^ we owe the estab- 

 lishment of the third sub-division, or that of the Grossopteri, dis- 

 tinguished by their "lobate" paired fins, the more or less diphy- 

 cercal (never homocercal) tail, the absence of branchiostegal rays, 

 the presence of jugular plates, etc. They form five families and 

 two " sub-series." 



a. Rhombiferi. 

 Ehombodipterines (2 dorsal fins, 



vonian and Carboniferous). 

 Folypterini (a multifid dorsal 



Eecent). 



b. Cycliferi. 

 De- 3. Cyclodipterini (2 dorsal fins, Devonian 



and Carboniferous). 

 fin. 4. Phaneropleurini (a single dorsal fin, 

 Devonian and Carboniferous).* 

 5. Ccdacanthini^ (Carboniferous, Per- 

 mian, Triassic, Liassic, Oolitic, Cre- 

 taceous). 



^ In Aspidorhynchus, wbere they are generally thought to be wanting, they are 

 nevertheless figured and described by Pictet. 



^ A single species from Lebanon, Falaobalistum Goedelii. 



3 Strange to say, the ingenious "preliminary essay" of this excellent author is 

 far less known among continental palaeontologists than it ought to be. No other work 

 since the time of A^assiz's •' Recherches " has advanced the progress of Palseichthy- 

 logy so much as this admirable book. "Will the author forgive my suggesting the 

 slight modification of his arrangement expressed in the terms '■'• Rhotyibodipterini" 

 and " Cyclodipterini'' ? or my intimating that Qyroptychius should apparently be 

 removed to the cycloid division of the JJipterini^ 



4 " Uronemus lobatus,'" Ag. and Gcelacanthus Miinsteri fFermian ?) are said to be 

 related to Fhaneropleuron. 



5 To the genera enumerated and discussed by Prof. Huxley in his recent monograph 



