EVOLUTION AND PALEONTOLOGY 253 



siders nevertheless that the latter represents the ancestral form 

 of the Acanthodians. 



PleuracantJius, which occurs in the Carboniferous and Per- 

 mian both in Europe and North America, on the other hand, 

 differs in many respects from sharks, and in these same features 

 approximates to the Dipnoi. It was a shark in the following 

 features : the gill-slits opened separately on the surface ; the 

 jaws consisted of the same primitive cartilages as in sharks, 

 claspers occurred in the male. On the other hand, the paired 

 fins show a transition to the feather-like or pinnate structure 

 which is in reality not primitive but secondary, the basal carti- 

 lages being separated from the body at the posterior end, and 

 the radial cartilages extending round to the posterior, originally 

 the internal, side, so that a structure is produced which has a 

 central axis and rays along each side. But the feature most 

 similar to the lung-fish was that the dermal denticles had dis- 

 appeared from the skin of the trunk, and on the head had de- 

 veloped into dermal bones arranged like those of the lung-fish. 

 The tail retained the primitive symmetry and was not of the 

 heterocercal or shark type. 



PleuracantJius, therefore, looks like the ancestor of the lung- 

 fishes, and seems to show how these were derived from the 

 shark type. 



True Dipnoi, however, are found in strata much older than 

 PleuracanthuS) and we can only suppose that the latter type 

 may have been in existence earlier than the Dipnoan, though 

 not preserved or not yet discovered. 



It might be expected, since the shark type is assumed to be 

 the earliest, that forms like the existing sharks and dog-fishes 

 would occur among the fossils of the oldest rocks, whereas the 

 examples we have mentioned are in many important characters 

 different from existing species of the group. The answer to 

 this is that the cartilaginous skeleton, would be likely to putrefy 

 before being imbedded in sediment, and that large numbers of 

 spines and teeth certainly belonging to fishes of the group are 

 preserved. Again it is only maintained that the type was the 

 earliest, not necessarily that all the existing forms have come 

 down from the earliest times unchanged. Of existing forms 

 Chlamydoselaclius (Plate XXI., B) which occurs in about 150 

 fathoms off the coast of Japan, off Madeira and off Norway and 



