504 



NATURE 



[September 20, 1900 



That Psamiiiosteus had an ancestry similar to that of Drepan- 

 aspis can also hardly be doubted. 



Finally, in the well-known Pteraspis of the Upper Silurian 

 and Lower Devonian formations we have a creature which also 

 has the head and anterior part of the body enveloped in a 

 carapace, to which a tail covered with rhombic scales is 

 appended behind, and, though the caudal fin has never been 

 properly seen, such remains of it as have occurred distinctly 

 indicate that it was heterocercal in its contour. The plates of 

 the carapace have a striking resemblance in general arrange- 

 ment to those of Drepanaspis, though the small polygonal 

 pieces have disappeared, and there is a prominent pointed 

 rostrum in front of the mouth ; and it is to be noted that the 

 small round apertures usually supposed to be orbits are in a 

 position quite analogous to that of the sensory pits in Drepan- 

 aspis. The plates of the carapace of Pteraspis are not, how- 

 ever, tuberculated, but ornamented by fine close parallel ridges, 

 the microscopic structure of which, along with their frequent 

 lateral crenulation, leaves no doubt in our minds that they have 

 been formed by the running together in lines of ThelodusAWi^ 

 shagreen grains. An aperture supposed to be branchial is seen 

 on the plate forming the posterior angle of the carapace on 

 each side. 



Until these recent discoveries concerning the Crslolepidae and 

 Drepanaspidse, Pteraspis and its allies, Cyathaspis and 

 Palaeaspis, constituted the only family included in the order 

 Heterostraci of the sub-class Ostracodermi, distinguished, as 

 shown by Lankester, by the absence of bone lacuna in the 

 microscopic structure of their plates. It is now, however, clear 

 that we can trace them back to an ancestral family in which the 

 external dermal armature was still in the generalised form of 

 separate shagreen grains or spinelets. 



But the Ostracodermi are usually made to include two other 

 groups or orders, namely the Osteostraci and the Asterolepida.^ 



The Osteostraci are distinguished from the Heterostraci by 

 the possession of lacunoe in their bone structure, and by 

 having the eyes in the middle of the head-shield instead of at 

 the sides. Cephalaspis, which occurs from the Upper Silurian 

 to the top of the Devonian, is the best known representative of 

 this division. Instead of a carapace, we find a large head -shield 

 of one piece, though its structure shows evidence of its having 

 been originally composed of a mosaic of small polygonal plates, 

 and it is also to be noted that the surface is ornamented by 

 small tubercles, there frequently being one larger in size in the 

 centre of each polygonal area. The posterior-external angles 

 of the shield project backward in a right and left pointed 

 process or cornu, scarcely developed in C. Murchisoni, internal 

 to which, and also organically connected with the head-shield, 

 is a rounded flap-like structure, which strongly reminds us of 

 the lateral flaps of the Coelolepidje. The body is covered with 

 scales, which on the sides are high and narrow ; there is a small 

 dorsal fin, and the caudal, though heterocercal, is not bilobate. 

 It is scarcely necessary for me to add that we find just a little 

 evidence of jaws or of teeth as in the case of the Heterostraci. 



The association of the Heterostraci and Osteostraci in one 

 sub-class of Ostracodermi has been strongly protested against 

 by Prof. Lankester and Dr. O. M. Reis, but here the Scottish 

 Silurian strata come to the rescue with a form which I described 

 last year under the name of Ateleaspis tessellala, and of which 

 some more perfect examples than those at my disposal at that 

 time have recently come to light through the labours of Mr. 

 Tait, of the Geological Survey of Scotland. 



Here we have a creature whose general form reminds us 

 strongly of Thelodus, but whose close affinity to Cephalaspis is 

 absolutely plain, were it only on account of the indications of 

 orbits on the top of the head. 



The expanded anterior part which here represents the head- 

 shield of Cephalaspis shows not the slightest trace of cornua, 

 but forms posteriorly a gently rounded lobe on each side, 

 clearly suggesting that the cornual flaps of Cephalaspis are 

 homologous with and derivable from the lateral expanses in the 

 Coelolepidse. This cephalic covering is composed of numerous 

 small polygonal plates like those of which the head-shield in 

 Cephalaspis no doubt originally consisted, and the minute 

 tubercles which cover their outer surfaces also suggest that the 

 superficial layer was formed by the fusion of Coelolepid scales. 



1 To these I myself recently added a fourth, the Anaspida, for the 

 remarkable Upper Silurian family of Birkeniidse, but as these throw no 

 light as yet on the problem of Descent they may at present be only 

 mentioned. 



NO. 1612, VOL. 62] 



The body is covered with rhombic scales, sculptured externally 

 with tubercles and wavy transverse ridges, and arranged in lines 

 having the same general direction as the scutes of Cephalaspis, 

 from which we may infer that the latter originated from the 

 fusion of scales of similar form. The fins are as in Cephalaspis, 

 there being one small dorsal situated far back, and a heteroceral 

 caudal, which is triangular in shape, and not deeply cleft into 

 upper and lower lobes as in the Coelolepidre. Finally, the 

 scales, on microscopic examination, show well-developed bone 

 lacuna' in their internal structure. 



That Ateleaspis belongs to the Osteostraci there is thus not 

 the smallest doubt, but its general resemblance to the Coelo- 

 lepida in its contour anteriorly led me to regard it as an annectent 

 form, and consequently to believe that there is after all a 

 genuine genetic connection between the Heterostraci and the 

 Osteostraci. And I have not seen reason to depart from that 

 opinion even th ugh Ateleaspis turns out to be still closer to 

 Cephalaspis than was apparent in the original specimens. 



If this be so, then Cephalaspis, as well as Pteraspis and its 

 allies, is traceable to the Ccelolepidge, shark-like creatures in 

 which, as we have already seen, the dermal covering consists of 

 small shagreen-like scales, or of minute hollow spines, and 

 consequently all theories as to the arthropod origin of the 

 Ostracodermi, so far as they are founded on the external con- 

 figuration of the carapace in the more specialised forms, must 

 fall to the ground. And from the close resemblance of these 

 scales of Thelodus to Elasmobranch shagreen bodies — for forty- 

 five years they had been, by most authors, actually referred to 

 the Selachii — I concluded that the Coelolepida owed their origin 

 to some form of primitive Elasmobranchs. That is, however, 

 not in accordance with the view of the late Prof Cope, that the 

 Ostracodermi are more related to the Marsipobranchii, and that, 

 from the apparent absence of lower jaw, they should be placed 

 along with the last-named group in a class of Agnatha, altogether 

 apart from the fishes proper. And Dr. Smith Woodward, who 

 is inclined to favour Cope's theory, has expressed his view that 

 the similarity of the Coelolepid scales to Elasmobranch shagreen 

 is no proof of an Elasmobranch derivation, but that such struc- 

 tures, representing the simplest form of dermal hard parts, may 

 have originated independently in far distant groups.^ Knowing 

 what we do of the occurrence of strange parallelisms in evolution, 

 it would not be safe to deny such a possibility. But as to a 

 Marsipobranch affinity, I would point out that the apparent 

 want of lower jaw among the hard parts which nature has pre- 

 served for us is no proof of the absence of a Meckelian cartilage 

 among the soft parts which are lost to us for ever ; and also, as 

 Prof. Lankester has remarked, that there is no evidence what- 

 ever that any of the creatures classed together as Ostracodermi 

 were monorhinal like the Lampreys. The only fossil vertebrate 

 having a single median opening, presumably nasal, in the front 

 of the head is Palaeospondylus, but, whatever be the true affinities 

 of this little creature, at present the subject of so much dispute, 

 I think we may be very sure that it is not an Oslracoderm. 



The Devonian "Antiarcha" or Asterolepida, of which 

 Pterichthys is the best known genus, are also usually placed in 

 the Ostracodermi, with which they agree in the possession of a 

 carapace of bony plates, in the absence of distinct lower jaw or 

 teeth, in the non-preservation of internal skeleton, and in haviiig 

 a scaly tail furnished with a heterocercal caudal fin, and, as in 

 the Cephalaspida, also with a small dorsal. But they have in 

 addition a pair of singular jointed thoracic limbs, evidently 

 organs of progression, which are totally unlike anything in the 

 Osteostraci or in the Heterostraci, or indeed in any other group 

 of fishes. These limbs are covered with bony plates and hollow 

 inside, but though I once fancifully compared them in that 

 respect with the limbs of insects, I must protest strongly against 

 this expression of mine being quoted in favour of the arthropod 

 theory of the derivation of the Vertebrata ! 



Nor do I think that there is any probability m the view pub- 

 lished by Simroth nine years ago,- namely, that Pterichthys may 

 have been a land animal which used its limbs for progression on 

 dry ground, and that the origin of the heterocercal tail was the 

 bending up of the extremity of the vertebral axis caused by its 

 being dragged behind the creature in the act of walking. That 

 view was promulgated before the discovery of the membranous 

 expanse of the caudal fin in this genus. 



But though the Asterolepida are apparently related to and in- 

 clusible in the Ostracodermi, the geological record is silent as 



1 Geol. Mag., March 1900. 



2 " Die Entstehung der Landthiere," Leipzig, 1891. 



