LINNEAN SOCIETY OF LONDON. ^;^ 



¥'me large specimens of Dinichtlnjs (figs. 4 C, 4 D) and its allies, 

 iiowever, are well pre.serv'ecl in the Upper Devonian flagstones of 

 Ohio, U.8.A., antl these are merely tiattein^d by crushing. I 

 have already pointed out that in Homostens the bony cranial shield 

 extends backwards beyond the brain *, and there can be no doubi- 

 thut in all Arthrodira it covers fhe branchial chambers (fig. 4 D, 

 hr.) behind the occiput. The so-called median and lateral 

 occipital elements, therefore, are not strictly cranial bones, but 

 really posterior to the cranial roof. The posterior end of the 

 brain-case is marked by a pair of ossified cartilages {ex.) which 

 seem to correspond with those named exoccipitals in the existing 

 Dipnoi. The upper part oF the lateral walls of the brain-case is 

 also ossified, firmly united with the cranial roof, and extends 

 backwards at each posterior angle to form part of the anterior 

 wall of the branchial chamber. The wide antorbital portion of 

 this ossification (ec<.), transversely grooved to lodge the small 

 anterior jaw-bone, may be regarded as the lateral ethmoid (or 

 ectethmoid). Immediately behind the orbits the ossification gives 

 rise to a pair of stout descending processes (.i'.), which are of un- 

 determined nature. There is no trace of any ossification in the 

 floor of the cranium, and no parasphenoid or other sheathing bone 

 has hitherto been observed. The nasal and otic capsules also have 

 not been seen, and there is no perforation in the cranial roof for 

 an aqueductus vestibuli. 



The bones of the cheek and jaw are well shown in their natural 

 relative positions in a specimen of Diiichthys in the British 

 ^Museum (no. P. 9340). The suspensorium, however, is as usual 

 not completely ossified, and there is onlv one isolated bone in each 

 ramus of the mandible (fig. 5,c?.). The single cheek-plate {so.) is 

 articulated with the cranial shield by a Vfry slight overlap, and 

 covers the whole of the postorbital region of the cheek ; it also 

 extends beneath the orbit as a narrow bar, from which an almost 

 semicircular flange extends downv\ards in contact with the outer 

 face of the supposed maxilla (fig 4 E). The cheek-plate is marked 

 by a groove for the circumorbital sensory canal, from whirh a 

 branch extends backwards close to and nearly parallel with the 

 lower border. The bone which occupies the position of a maxilla 

 Cmx.), and may perhaps be so identified, is a very stout blade, 

 longer than deep, with a sharp cutting edge at the oral border. 

 It is firmly adpressed to the flange of the cheek-plate already 

 mentioned, and at its upper border near the anterior end it bears 

 a small but stout inwardly directed process (fig. 5, })r.). An 

 anterior bone {pmx.) which seems to occupy the position of a pre- 

 maxilla (and in some species is tiiberculated on the outer face), is 

 produced backwards at its upper end in a process which overlaps 

 the process of the supposed maxilla, and is interposed between 

 this and the lateral ethmoid of the skull, into a groove of which 

 (fig. 5, ect.) the whole ujiper edge oF the bone fits. Tliis supposed 

 premaxilla slightly overlaps the inner face of the anterior end of 



* A. S. Woodward, Proc. Zool. Soc. 1891, pp. 198-201. 

 LINN. SOC. PJU)CEEDINGS. — SESSION 1921-22. d 



