332 Dr. C. Liitken on the Limits and 



guisli between the genera with large and those with small 

 scales, and between the heterocercal and subhomocercal types ; 

 and in this way we should obtain a quaternary division such 

 as this : — 



1. Lepidosteidfe lieterocercae microlepidotse Cheirolcpis. 



2. ,, liomocercaB ,, „ Saurojms. 



3. „ heterocercae macrolepidotse .... Palceonkcus. 



4. ,, homocercae „ „ Lepkhtus. 



But it appears to me to be impossible to mark out fixed 

 limits between these groups, which are artificial rather than 

 natural. It has also been proposed to divide the Eugano'idei 

 into " monostichi " and " distichi," accordino; to the sinsrle 

 or double arrangement of the scales bordering the fins ; but 

 we are still destitute of sufficient information to enable us to 

 adopt this classification, even if it has an actual foundation in 

 nature. 



Eveiy one knows that there is a difference of epoch between 

 tlie Eugano'idei called " heterocercal" and tliose called " homo- 

 cereal," or, better, " simorrhachal ;" but the line of demarca- 

 tion is not so clearly drawn as has been supposed. As early 

 as the Permian system there are species (refeiTed to the genus 

 Palceoniscus) which are only semiheterocercal, whilst in the 

 Lias we may still find absolutely heterocercal genera {Oxy- 

 gnathus, CosmoJejns). In general, however, an evident pro- 

 gress from the heterocercal to the so-called homocercal or fan- 

 like tail may be observed running parallel to the progress of 

 geological epochs. A similar progress is marked also, al- 

 though perhaps less distinctly, in the structure of the vertebral 

 column. No Lepidosteid presents true biconcave vertebral 

 bodies : except in the living Lepidostei^ we find either a naked 

 nofocJwrd without any trace of vertebral bodies, the apophyses 

 of the vertebrae, the interapophysial bones, the scapular arch, 

 the fin-rays, &c. being at the same time well developed and 

 ossified; or semiverfebrcp^ that is to say, superficial plates, de- 

 rived from the neurapophyses and ha3mapophyses, covering 

 the notochord completely or partially, and frequently, by 

 touching or covering each other, simulating false vertebrte ; 

 or, lastly, these plates becoming amalgamated, so-called 

 anmdar vertehro'j differing, however, from the true vertebrje 

 of fishes by their smooth surface and their bony interior en- 

 closing the notocliord, almost completely developed. The 

 reader who may Avish to have more ample information upon 

 this subject I recommend to consult especially the works of 

 MM. Heckel and Wagner. 



II, Second series. — The Lpj^idoj^leuridce or Pycnodontes are 



