No. 436.] APPENDAGES OE TREMA TASPIS. 



22 9 



its asymmetry is clearly seen. Unfortunately the characteristic 

 texture of the median surface is not well shown in the photo- 

 graph. This was due in part to the fact that the articulating 

 condyle was accidentally broken off and the shim surface of the 

 cement used to repair it made it difficult to obtain good photo- 

 graphs. It shows, however, with sufficient distinctness that the 

 enamel layer which covers only the posterior end of the median 

 surface, abruptly terminates anteriorly in a thick fold. mc'. 



The remaining portion of the median surface is covered with 

 a peculiar layer entirely different in texture from that on the dor- 

 sal and ventral surfaces. It is much darker colored and its lus- 

 treless surface, which is irregularly folded, is devoid of pores, 

 except a few scattering ones of comparatively large size. It 

 has the appearance of having been a tough flexible membrane 

 that had become dried and wrinkled before fossilization. Nearer 

 the anterior end, the conspicuous median ridge and the well 

 moulded condyle indicate the presence of firmer bone-like 



I regard the whole structure as the distal joint of an oar-like 

 swimming appendage that served the same function tor Trcm- 



Pterichthys. That we are dealing with a projecting appendage 



We know with certainty that such a plate could not have 

 been a mere spine attached to either the ends of the cornua, or 

 to the lateral margins of the cephalic buckler back of the oral 

 region, or to any part of the dorsal surface of the head, for on 

 these parts, the structure of which is perfectly well known, there 

 are certainly no articulating surfaces that could serve for the 

 attachment of either spine's or appendages of this character. 



sal or ventral spine of the trunk, for the difference in texture 

 and contour of its two surfaces proves that it was a part, or the 

 whole, of a paired organ. We are, therefore, limited to the 



