CEPHALASPIDiE. 



3.-) 



continuation of the superficial marking to the inferior surface is, of course, simply that 

 the parts of the inferior aspect of the shield so marked were as anatomically superficial 

 as the tuberculatcd superior surfiicc. The recognition of this fixct is im{)ortant, as being 

 inconsistent with the supposition of a mouth reaching along the margin of the fish's head, 

 which involves the notion that the upper jaw is formed, as it were, by the margin. Probably 

 no one acquainted with recent forms allied to the Ceplialaspida would hazard such a 

 conjecture. The woodcut, fig. 15, is interesting to compare in the consideration of this 

 matter. The tubercles on the margin of the shield sometimes assume a very spinous or 

 even tooth-like form, e.(/. in the species C. asper and Ealcerasjns pastulifer (see woodcut, 

 fig. 23), and hence it is not surprising that detached portions of the margin with its 

 tubercles should have been mistaken for bits of fish-jaw. On the inner side of the 

 cornua the tubercles are modified ; even the common Herefordshire species presents 

 the form of long tooth-like excrescences, and in C. asper, one of Mr. Powrie's disco- 

 veries, the scales, which are always tuberculatcd in the Osteostraci like the head-shield, 

 present very close-set and acute spinous ornamentation (PI. X, fig. 5). 



The tubercles themselves are generally round, with a bright and smooth surface, 

 similar to that of the intervening matter, both resembling in their microscopic structure 

 the * cosmine ' of Prof. Williamson. In one species discovered by Mr. Lightbody, of 

 Ludlow, in the beds of the Tilestone group which have furnished the Auchenasjns 

 SaKeri, the tubercles have a curious truncated appearance, giving them a really crater- 

 like form. In other parts of the same head-shield, or on what are perhaps additional 

 posterior plates, the tubercles are intersected by a network of coarse grooves forming 

 an irregular crocodiloid pattern. The tubercles of Auchenaspis are large in proportion 

 to the size of the shield ; those of EuJceraspis very closely set and small, though not so 

 small relatively as those of Cephalaspis Powrici and Cephalaspis A(jcmizii. 



The exceeding rarity of the i)rcservation of the tuberculatcd surface in specimens of 

 Osteostraci must be borne in mind. I have only one specimen of the common Here- 

 fordshire species that shows them at all well. It is nmch more diflicult to get this 

 surface in a state of preservation than that of the Pteraspids, difficult as that is to obtain 

 Frequently the lower layers of the Osteostracous shield are preserved, and have been 

 mistaken for the true superficies by Prof. Agassiz and by many collectors. When the 

 matrix is of such a character as to preserve well the test, the shield is very generally 

 crushed and distorted, so that we are almost invariably offered a choice of evils. 



There are no ' pits' on the surface of the Osteostracous shields analogous to those 

 described by me in the genera Scap/taspis and Pteraspis, which seem to represent the 

 lateral line. 



The scales, like the head-shields in Ccphalaspidce, are to be looked upon as bones or 

 parostotic formations, and not as mere scales. They belong to the aponeurotic region 

 spoken of by ;Mr. Parker in his great work,^ and, the ganoin-layer being absent from them, 

 1 'On tlie Shouldcr-girdl.,' &c. (Roy. Soc), tto, 1S48. 



