40 WATSON & Day, Notes on some Palceozoic Fishes. 



important a point to be overlooked. We have therefore 

 shown that so far as concerns the dorsal surface of its 

 head Ceratodus is a morphological descendant and very 

 probably a real descendant of Dipterus. Dollo has 

 already shown the same fact so far as concerns the 

 median fins. 



Passing on to other independent structures we find 

 that in side view there is a very striking resemblance 

 between the heads of Ceratodus and Phaneropleuron^ the 

 differences being mainly due to the further reduction of 

 the circumorbital and other bones. PJianeropleuron, how- 

 ever, is in this region as in others obviously derived by 

 deficient ossification or by reduction from Scaumenacia 

 which itself has undeniable resemblances to Dipterus. It 

 will be noticed that in all these genera the ridges of the 

 dental plates radiate from a point, and that there seems 

 to be a gradual reduction from twelve ridges in the 

 palatine tooth of Dipterus platyceplialus, to seven in Scau- 

 menacia, six (seven ?) in PJianeropleuron, six in Sagcnodus, 

 five in Ceratodus ornatus (Middle Trias, South Africa), 

 four and a rudiment in Ceratodus latissimus (Rhaetic) 

 again suggesting that we have to do with a real genetic 

 series. Ctenodus, whose skull roof has not at present 

 entered into our discussion, differs from the members of 

 this series in its dentition, in the fact that the ridges on 

 the teeth are sensibly parallel, there being from twelve to 

 fourteen ridges in Cteuodus interrupt us from the Lower 

 Carboniferous of Scotland, twelve to fourteen in Ctenodus 

 cristatus from the Middle Coal Measures, and twenty in 

 Ctenodus murcJiisoni from the Upper Coal Measures. It 

 thus apparently differs from the members of our series 

 not only in the parallelism of the ridges but also in their 

 tendency to increase in number. 



If the diagram of the top of the head {Text-fig: 8, H) 



