24 SIRENOID AND CROSSOPTERYGIAN GANOIDS. 



important differences between Ceratodus and Dipterus, of which perhaps the most 

 obvious are the diphycercal tail of the first and the heterocercal tail of the second ; but 

 we cannot entirely cut the two genera off from each other. Nor can we, I think, admit 

 that the Ganoidei, Plagiostomi, and Teleostei have had a common ancestor from which 

 Ceratodus and Lepidosiren derive no lineage. It remains for future investigation to 

 draw the lines of connection and separation with their exact strength as well as in their 

 true place. Prom such data as are now accessible it would appear that, on the whole, 

 the Sirenoidei are more widely separated from other autostylic Ichthyopsida (such as 

 Amphibia and Chimcera) than from the Crossopterygian Ganoids. 1 



In defence of this position it may be urged that another adaptation at least equally 

 special with the displacement of the suspensorium, viz. the very peculiar dentition, 

 connects together the four genera Ceratodus, Lepidosiren, Dipterus, and Ctenodus, while 

 these are associated, not only with one another but also with the acute-lobate Crosso- 

 pterygian Ganoids, by the structure of their paired fins. This last feature of resem- 

 blance, unimportant enough on the physiological side, seems to have no inconsiderable 

 classificatory importance ; and though the presence or absence of scales upon the fin does 

 not mean much, yet the acute-lobed patch seems as yet to connect together only fishes 

 between which some close affinity may be supposed to exist. 



The arguments for and against the near alliance of the Dipnoi with Dipterus and 

 Ctenodus are summarised in the following table : — 



Paired fins acute-lobate, fringed; 3 den 

 tition of opposed pairs of radiately 



A. 



Lepidosiren j 



Protopterus r Autostylic ; no gular plates ; archi- 



ridged plates and an unopposed pair-^ Ceratodus ) pterygium ; dipbycercal. 



of vomerine teeth ; nasal passages 

 opening into the mouth [Ctenodus ?] ; 

 no ossified branchiostegal rays. 



B. 

 Dipterus ... ") Hyostylic ? ; gular plates ; archiptery- 

 Ctenodus...) gium ? ; heterocercal {Dipterus). 



It is not known whether Dipterus and Ctenodus are hyostylic or autostylic ; but 

 Pofypterus is hyostylic. It is not known for certain whether Dipterus and Ctenodus 

 have an archipterygium or an ichthyopterygium ; but the outward resemblance of the fins 

 of Dipterus to those of Ceratodus renders it probable that they have an archipterygium. 



1 Prof. Huxley's table ('P. Z. S.,' 1876, p. 59) is apparently not intended to comprehend all the 

 data essential to a true view of the affinities subsisting between different Ichthyopsida. It is introduced 

 (p. 58) as a "mode of stating the facts of morphology in a condensed and comprehensible form, which 

 shall be purely objective and free from speculation." 



2 Lepidosiren and Protopterus can only come under this definition by a considerable stretching of 

 the words. It has already been explained what the structure of the paired fins in these genera really is. 

 The analogy between the paired fins of Protopterini and those of the acute-lobate Crossopterygii was 

 pointed out by Prof. Huxley in his ' Essay on the Classification of Devonian Fishes,' p. 26 (1861). The 

 fins of Ctenodus are not known. In Lepidosiren the paired fins are not fringed. 



