O.sfracodntii fnim ('nii'iiit. 417 



sido-virw. 'I'lu' (Ifiino-skclrtiil covorinpj of tho lu-ad must 

 have hi (Ml very -sli^lit, tor its border cannot 1)(» distiii^^uislied 

 witli certainty. It seeni.s to Ijave consisted njerdy of small 

 iilia;;reen-like ;,'ranule.s, wliicli arc not fused into plates. The 

 only clearly distinj^uishable feature is a pair of small thick- 

 ened skeletal rings (o.), which may probably be interpreted as 

 orbits, but are somewhat further apart tlian in Cephalaspis. 

 The shagreen-like granules are seen within each supposed 

 orbit ; but this may be the result of accidental disjjlacement 

 or the crushing of the dorsal shield upon the ventral c<jvering. 

 ^V'hen viewed in sonie lights other a|)pearance3 suggest that 

 more structure might be observable in a better preserved 

 specimen; but a slightly fibrous longitudinal streak appa- 

 rently beneath the outer granules, extending backwards troni 

 the left orbit-like ring, is the only other noteworthy feature. 



The abdominal region is probably deepened by the distor- 

 tion of its dorsal portion, while its dermal covering is nearly 

 as obscure as that of the head. It seems, however, to have 

 been armoured with small, narrow, and deep scales, arranged 

 in straight rows, which incline slightly forwards and down- 

 wards instead of backwards and downwards, as is usually the 

 arrangement in fishes. At the anterior end of the ventral 

 border there are obscure traces of two or three slightly larger 

 scales, which may have been placed on the median ventral 

 ridge. The less complete side of the fossil (PI. X. fig. la) 

 shows not only the peculiar squamation, but also certain rod- 

 like remains near the dorsal border of the abdominal region, 

 which are suggestive of calcified neural spines of an endo- 

 skeletal axis. The latter structures, however, are not quite 

 cle.ir. 



There are no traces of paired fins or their supports. 



The slender caudal region is well shown in direct side-view 

 on both halves of the fossil, and terminates in a heterocercal 

 tail. It is completely covered with scales like tliose of the 

 abdominal region, similarly disposed in straight oblique series 

 inclined forwards and downwards. The scales seem to be 

 rather thick, scarcely if at all overlapping, invested with 

 enamel and marked with a few antero-posteriorly-directed 

 ridges and grooves. One from the middle of the flank is 

 represented diagrammatically of ten times the natural size 

 in tig. 1 b. The scales on tlie sides of the upper caudal lobe 

 seem to be nearly quadrilateral, while its dorsal border is 

 fringed with a single (possibly paired) series of very slender 

 ridge- scales. There is a small remote dorsal fin (r/.), low and 

 triangular, ajiparently stiffened by scales like those of the 



Ann. d; Ma;j. S. Hint. Ser. 7. Vol. v. 27 



