36 



OLD RED SANDSTONE EISHES. 



there may have existed a separate overlying set of scales, though the ornament renders 

 this unlikely. 



Of the Form and Construction of the Shield. — In the Osteostraci the shield is very 

 greatly convex when not distorted by pressure, attaining an almost hemispherical 

 curvature in some parts. The test bends over at the margin, and is continued inwards, 

 forming a hollow rim to the shield, which at its posterior angles widens out, as seen in 

 the woodcut (fig. 12), forming what may be termed the 'lateral floors' of the cephalic 

 shield. Since these inflections terminate abruptly, there is no (general floor to the cephalic 

 box as far as calcareous matter goes. It was completed by a tough skin and a mouth, no 

 doubt. The test in Cejjhalaspis is, moreover, reflected from the margin and floor against 

 the roof of the shield, forming a lamina of great tenuity subjacent to the upper tuberculated 

 lamina which forms the roof of the shield. This secondary 'roof'i thins out towards the 

 centre of the cephalic plate, and probably became membranous (see woodcut, fig. 13, section ; 

 and M'oodcut, fig. 12, inferior aspect). The hollow rim must be compared with the 

 hollow cornua oi Pterasj)is ; and probably like them it had apertures at its posterior angles, 

 close to the attachment of the pectoral appendages (see fig. 16). In the genus Cephalaspis 

 proper, which may be placed subgenerically as Eucephalaspis, the shield is ellipsoid in 

 its anterior portion, and is produced into cornua of considerable length at each posterior 

 angle. The median portion is also extended in a backward direction, and in the middle 

 line presents a short spine. In the subgenus Hemicyclaspis the cornua are not developed, 

 nor is the median portion markedly produced, the whole shield having a truncated 

 hemispherical outline. In the subgenus Zenaspis the cephalic shield has the character 

 of that of Eucephalaspis ; but one at least, and perhaps more, scutes placed dorsally in the 

 median line succeeded it. This arrangement paves the way to Sir Philip Egerton's 

 genus Auchenaspis, in which the shield is identical with that of Eucep)halaspis, supposing 

 its median posterior portion but slightly developed, and a large neck-plate soldered to it. 

 The subgenus Euheraspis, which I have associated with Auchenaspis, though no neck- 

 plate has been yet found attached to the cephahc portion, possessed exceedingly long 

 cornua, extending backwards to more than double the length of the shield itself. Other 

 characters of importance render it desirable to form this subgenus of Auchenaspis. The 

 genus Didi/niasjns had a neck-plate fully as large as the anterior or cephalic plate, which 

 appears to have possessed no cornua. The neck-plate may perhaps be regarded as 

 representing in those genera which possess it the distal portion of the compound 

 shield of Fterasjris, though it is not advisable to trace a very close morphological 

 relation between the two groups Heterostraci and Osteostraci in the details of the 

 construction of their superficial armature. The remarks which were made (p. 17) with 

 regard to the nature of the suturing or fusion of the pieces of the shield in the Pteraspids 

 apply to the similar cases among Osteostraci with equal force, the dift'erence of histological 

 composition in the two cases being, however, duly taken into consideration. 



The position of the orbits, and certain very remarkable concavities and eminences 

 ' Observed by Prof. Huxley. 



