T.TXXE.VX SOCIETY OF LONDOX. 3 I 



beginning immediately behind the dorsal spine already mentioned, 

 while the corresponding ventral series of fnlcral scales stops at a 

 short distance rroin the iiinder border of the median ventral plate, 

 leaving a gap doubtless for the cloaeal opening. 



When Traquuir * ilrst interjjreted Dtrpanaspis he reversed the 

 dorsal and ventral aspects as just described, l)ecause he observed 

 that the stronger and longer lobe of the tail was apparently on 

 the same aspect as the single median plate, while the cleft in the 

 liinder of the two opposing median plates might be regarded as 

 the cloaeal opening. The inter]->retation of Deanf and Kia^r;]:, 

 however, now adopted, seems to be confirmed by the recent dis- 

 covery of a nearly complete specimen of PhylJolepis (fig. 3) §, 

 which is evidently an Upper Devonian representative of the same 

 group. Here the dorsal aspect is undonbtedly that on which the 

 two median i)lates occiir, and the ventral aspect bears only one 

 large plate (perhaps paired) with an extensive gap in front for 

 the mouth-parts. In front and round part of the sides of 

 the median dorsal plates there is a single row of marginal plates, 

 ending behind in a pair of backwardly-directed coruua ; but the 

 orbits do not pierce any of these plates, and their exact position, 

 though evidently lateral, is uncertain. The small tail is scaleless, 

 and in tlie fossil it shows remains of the superficially calcified 

 neural and liosmal arches of the notochordal axis. 



OSTEOSTRACI. 



Another order of Ostracoderms with a large depressed head is 

 armoured in this region with small polygonal plates, which are 

 variously fused together. Bone-cells are conspicuous in all the 

 hard tissue except the superficial layer of tubercles. The eyes 

 are close together in the middle of the top of the head. The tail, 

 as a rule, is relatively larger than that of the Heterostraci, ending 

 in the usual heterocercal fin ; and it is covered with scales, w'hich 

 are more or less deepened on the flank. The Osteostraci range 

 from the Upper .Silurian to the Upper Devonian, but are very 

 rare above the Lower Devonian. 



The armour of the typical widely-ranging genus Cepliahtapls 

 (fig. 4)11 bears several indications of the underlying soft parts. 

 The litrle plate between the orbits is marked by a pit, evidently 

 for the pineal body; and in front of it there are otlier markings 



* K. II. Traqiuiii-, Trans. Rov. Soc. Edinb. vol. xxxix. (lSi)9), p. S44, text- 

 fig. 5; loc. cit. vol. xl. (l'JO'3), p. 725, pi. 1. (Igs. l~'.l pis. 2-4, text-fi<rs, 1-3- 

 loc. cif. vol. xli. (190.^), p. 4G9, pis. 1-3; Gool. Ma^. [4] vol. vii.°(l<)00)' 

 p. 158, figs. 1-3 ; loc. cit. vol. ix. (1!)02), p. 289, figs. 1, 2. 



t B. Dean, Science, n. s. vol. xix. (1904), p. 64. 



{ J. Kiar, Rep. 2nd Norweg. Arctic Exped. ' Frani ' 1898-1902, no. 33 (1915) 

 pp. 29, 33, pi. 3. ligs. 5. 0, pi. 4. iig. 2, text-figs. G, 7. 



§ A. 8. Woodward, Rep. Brit. Assoc. (Australia, 1914), 1915, p. 122, pi. 2. 



11 See especially E. Ray Lankester in Powrie & Lankester, ' Fisiies' of tiie 

 Old Ked Sandstone,' pt. i. concluded (1870), and A. S. Woodward ' Catal 

 Enss. Fishes MM.: pt. ii. (1891), pp. 177-19.3, with (i<rg. 



