760 MR. E. S. GOODRICH ON [JSTov. 12, 



was put forward by "Williamson to account for the structure of the 

 scales and dermal plates of MegcdicJithys (26). Each pulp-cavity 

 •of the cosmine layer is supposed to indicate a single denticle. As 

 ■representing an intermediate stage in the foi-mation of cosmine, 

 he pointed to the dermal bones of the Ocelacanth Macropoma. 

 Here the denticles are clearly seen, either standing up as sharp 

 teeth on the scales, or as more or less blunt protuberances on 

 the cranial bones. The denticles in Macropoma i-etain their 

 histological character and pulp-cavity, although they become fused 

 to the underlying bone, and so deeply sunk in it, that they may 

 be mistaken for a mere superficial ornamentation. Cosmine, 

 •says Williamson, is formed by "the confluent aggregation and 

 superficial depression of a number of placoid teeth, surmounting 

 a highly developed scale." 



"While cosmine grows from ^vithin, ganoine is deposited from 

 without. The scale here called Pala?oni&coid, "Williamson believed 

 to be formed by the overlapping of the bony laminfe at the edges, 

 and their spreading out over the outer surface of the cosmine as an 

 enarael-like layer. The ganoid scale, then, would be described as 

 a cosmoid scale the edges of which have been turned over on to 

 the outer surface. 



These fascinating theories meet with many difiiculties in the way 

 of their adoption ; difficulties which are probably not insuperable, 

 iDut must be disposed of before "VV^illiam son's views can be con- 

 sidered as established. 



First of all, it is a far cry from the scattered denticles of 

 Macropoma to the complex cosmine of the Osteolepids. Most of 

 "Williamson's ubservations on Macropoma I can fully confirm. 

 But I have looked in vain for any intermediate forms bridging 

 over the gap between the two types of scale. The dentine-like 

 substance of Megalichthys forms a continuous layer, without sign of 

 subdivision externallj^ ; and the appearance of a series of crowded 

 ■denticles seen in a section is chiefly due to the pulp-cavities and 

 conical chambers being cut through alternately. Although so 

 extraordinarily like the shield of Fteraspis in section, the two are 

 really very different in structure. In the palasoniscoid scale the 

 cosmine may be said to have lost any trace of the subdivision into 

 separate denticles which it may once have possessed ; and in the 

 lepidosteoid scale no true cosmine occui-s at all, according to my 

 observations. 



Pander, it is true, has given beautiful figures of sections of the 

 dermal bones of Glyptolepis, showing how the superficial teeth 

 may become fixed to the bone, and converted into cosmine-like 

 tubercles (15) ; but I have failed to find anything quite like this 

 in the material at my disposal. On the scales of Glyptolepis 

 there appear to be no regularly disposed denticles, though 

 here and there ai'e tubercles which resemble rather the last 

 remnants of disappearing cosmine than the incipient stages in its 

 formation. Rohon, however, agrees with Pander, and supports 



