32 JOHAN KLER. [SEC. ARCT. EXP. FRAM 



structure is unfortunately unavailable". In this all subsequent investigators 

 who have dealt with this subject, are unanimous. 



I have now succeeded in obtaining the proof by reference to the 

 microscopic structure. The latter has been found to agree entirely on 

 the one hand with the structure of Psanimosteus, and on the other to 

 resemble to a great extent that of Pteraspis. In the case of the latter 

 the difference only consists in the circumstance that the vascular canal 

 system is composed of more regular lacunes, and that the denticles are 

 welded together into longitudinal ribs. 



This agreement with Pteraspidae, which is thus apparent in the 

 structure of the body plates, is even more evident when we consider the 

 form and arrangement of the body plates. 



We can hardly be in doubt as to how the large body plates may 

 be identified in these two types. That side of Drepanaspis which was 

 identified by TRAQUAIR as the ventral, in the median line has 2 large 

 plates following each other. (TRAQUAIR'S median ventral and mental 

 plate). At the sides we see the long cornual plates, and in front of 

 these the 2 pierced ocular plates. These quite correspond to the plates 

 on the dorsal side of Pteraspis, with its median dorsal plate, rostrum, 

 two cornua and two ocular plates. On the other side, which is identified 

 by TRAOUAIR as the dorsal, the large oval plate which TRAQUAIR described 

 as the median dorsal plate, agrees perfectly with the large ventral plate 

 of Pteraspis. The said plate agrees in the case of the two families, 

 not only in its arrangement but also in its form. The sole divergence 

 is found in the short form of the rostral plate, which however is quite 

 natural in the light of the ray-like body form of Drepanaspis. 



In consequence, it would be exceedingly remarkable if 

 the sides of the body, which in the case of these so closely 

 related species are characterised by the same development 

 and arrangement of plates, should not be identical. 



Of course an investigator like TRANQUAIR has not overlooked this 

 circumstance, but nevertheless he arrived at the opposite conclusion. 



The most important reasons for this are summed up by him in his 

 .supplemental work 1 on Drepanaspis as follows: - 



,,I submit therefore, that I have amply shown - 



First: - That the aboral aspect of the carapace of Drepanaspis 

 is coincident with the apparent dorsal "lobe" or aspect of the caudal fin. 



1 Supplement to the Lower Devonian Fishes of Gemunden, page 473. 



