GREGORY: FISH SKULLS 101 



of the head and body was strongly developed. Among the Permian pleuracanth sharks, 

 which were highly specialized for swamp life, the endocranium was strongly built. Cope 

 believed that in life it was subdivided into segments — whence the name Ichthyotomi; but 

 examination of Cope's material shows that the supposed sutures are merely fractures of the 

 fossilized mass, and that the skull of Diacranodus is essentially identical with that of modern 

 elasmobranchs. 



The skulls of modern sharks and rays are dealt with from a systematic viewpoint in the 

 works of Gegenbaur (1872), Tate Regan (1906a), Garman (1913), while the embryology of 

 the skull has engaged the attention of many authors from Balfour (1876-1878) to Sewert- 

 zofF (1916, 1917, 1927). The cranial anatomy of Chlamydoselachus is fully described in a 

 beautiful monograph by Allis (1923a). Many morphological problems of the elasmobranch 

 head are admirably discussed in the recent work of Goodrich (1930) and De Beer (1928). 

 The outstanding differences between the elasmobranch endocranium and that of primitive 

 teleostomous fishes will be summarized below (p. 115), but it may be said here that in 

 general the shark skull, besides lacking bony centers and derm bones, reflects the greater 

 development of the olfactory as compared with the optic parts of the skull; that it has not 

 yet developed either a parasphenoid or opercular bones; and that its hyomandibular is 

 attached below the vena capitis lateralis instead of above it (Goodrich). The palatoquad- 

 rate has typically two main points of attachment with the skull by means of palatobasal and 

 otic processes, whereas in the typical teleosts the posterior contact is lost and the quadrate 

 is suspended by the hyomandibular. The typical shark skull shows an early stage in 

 differentiation of the trigemino-faclalis chamber, a region of considerable morphological im- 

 portance in the higher types (Allis, 1914). 



The general form and details of the neurocranium of sharks and rays are naturally 

 influenced by many factors both external and internal. The elasmobranch skull is usually 

 wider and flatter than the teleost skull since the head as a whole is typically of the same 

 proportions. For various reasons it seems likely that the primitive chordates were not 

 swift-swimming, pelagic types but partly depressed, partly bottom-living forms (Gregory, 

 1928, pp. 389, 416), which at various times gave rise on the one hand to specialized bottom- 

 living types such as the cephalaspids, antiarchs and the rays, and on the other to fusiform, 

 partly free-swimming types, including the anaspid ostracoderms, macropetalichthyids, 

 coccosteids and earliest acanthodian sharks. The curious Devonian elasmobranch Gemun- 

 dina, as fully described by Broili (1930), perhaps stands midway in its habitus between 

 primitive, partly benthonic elasmobranchs and secondarily free-swimming ones. In one 

 direction a further broadening of its already broadened pectorals, without reduction of the 

 trunk or flanks, would produce a torpedo-like type, while a narrowing of its pectorals would 

 give rise to a Cestracion-Vike type. Further differentiation in opposite directions would 

 lead to the flattened eagle-rays and skates on one side, and to the swift pelagic mackerel 

 sharks on the other. But nowhere in the elasmobranchs do we find indications that the 

 remote ancestral stock were compressed, deep-bodied forms like the primitive spiny-finned 

 teleosts. Hence the elasmobranch occiput is never covered on top by extensions of the 

 dorsal body- or fin-muscles, the interorbital space is wide and the rostrum more or less 

 broad and depressed. This depression and widening, pronounced even in Mustehu, be- 

 comes emphasized in the bonnet-shark {Sphyrna tiburo) and excessive in the hammer-head 

 {Zygesna), where extreme transverse growth of the front end of the head has pushed the 



