34 



OLD RED SANDSTONE FISHES. 



depended on. But this is a very unsafe criterion, since, owing to tlie fact of its being 

 xeij thin, the shield is liable to the most deceptive distortion. Examples of this are 

 familiar to all who know the fish-heads of the Herefordshire Cornstones. So great is the 

 uncertainty in this matter, that I have had much doubt as to the distinction between the 

 Scotch type-specimen of C. Lyellii in the British Museum and another Scotch specimen 

 in the museum at Arbroath, and again between these two and the common Herefordshire 

 species included by Agassiz under Cej)7ialaspis Lyellii. 



Before proceeding to the description of genera and species, in Avhich I shall rely more 

 on the figures than on the words wdiich can be used to describe them, remembering the 

 motto inscribed by the illustrious Delia Chiaje in one of his works, " Res non verba," a 

 few general remarks may be oflTered with regard to the variations of the shield in 

 structure and form, and also a general description of the body-scales and fins. 



1. Variations in the Intimate Structure. — Eor a general description of the minute 

 structure of the test in the Osteostraci I must refer back to page 7 of this Monograph, 

 and also to the plate of microscopic drawings and explanation accompanying this part. 

 The microscopic structure of Cephalaspis Lyellii and of the large form called Ceplialaspis 

 asterolepis by Dr. Harley^ has been carefully studied, bat of other species proper material 

 has not come to hand. In Anchenaspis and Didymastpis I was able to ascertain the 

 presence of bone-lacunae of an elongated character, with their long axes crossing in the 

 various layers, thus giving the appearance of cross-hatching described by Prof. Huxley. 

 In the so-called Steganodidyum Carteri to w^hich I called attention in the ' Quart. Journal 

 of the Geol. Soc.,' 1868, p. 547, as being really an Osteostracous Cephalaspid (the 

 Steganodidyum Cornubicum having been pronounced a Pteraspid by Prof. Huxley and 

 Mr. Salter), I have found small bone-lacunae with very numerous fine canaliculi interlacing 

 in every direction. The lacunae w^ere not elongated, and were smaller in size than those 

 in the low^er portion of the shield of Cephalaspis Lyellii. 



2. Variations in the Surface-marldngs. — The tubercular ornamentation wdiich appears 

 to 1)0 a constant characteristic of this section of the ancient Sturgeons presents con- 

 siderable diversity of arrangement. In some cases each hexagonal area (described at 

 page 8, and figured in PI. XIV, fig. 7), supports a simple tubercle or one large 

 Tubercle and a smaller one. In other cases there are many of these tubercles, of a 

 smaller relative size, placed irregularly on the area ; or, again, there may be a central 

 tubercle surrounded by six or more smaller ones, like a miniature volcano w-ith its 

 secondary cones. At the margins of the shield, and especially on the 'cornua,' the 

 tubercles become very closely packed and elongated. The tubercular ornamentation is 

 continued on to the under surface of the shield, to some extent, along the enlarged 

 margin and its posterior expansions (see woodcut, fig. 12), just as in Fteraspis the 

 striations of the surface were shov^^n to continue on part of the inferior surface of the 

 rostrum, and along the margin formed by the hollow cornua. The meaning of this 



^ This species is not to be distinguished from the CepJialuspis Sahvey i Egcrton. 



