12 PROCEEDINGS OF THE NATIONAL MUSEUM vol. los 



In the great majority of the plates and scales, on the other hand, 

 the early shagreen of starlike tubercles has, during subsequent stages 

 of growth, been supplanted completely by another, mature type of 

 ornamentation consisting only of thick, mushroom-like tubercles (fig. 

 2,6; tf,, pi. 1, figs. 1, 2, 4), frequently of large size, which are in several 

 respects different from those on the exoskeleton of Astraspis. In top 

 view the crown of these tubercles is rounded, elongated, or even some- 

 what Iddney-shaped; it is always smooth and shining superficially, 

 and at its basal circumference it is generallj^ more or less crenulated. 

 The neck-portion of the tubercles, which is always well developed, is 

 a little constricted immediately below the crown, and is separated 

 from the latter by a distinct nick. Contrary to what might be ex- 

 pected in view of the state of preservation of the material, the smooth- 

 ness of the crown is certainly a primary condition. As in Astraspis 

 (Bryant, 1936, p. 418, pi. 2, fig. 1), the tubercles were in certain cases 

 subject to some post-mortem abrasion, but, as far as I can find, this 

 cannot possibly account for the fact that in tubercles of different 

 height, situated close together, and sometimes even partly superim- 

 posed on each other, the crown invariably shows the same smoothly 

 rounded contour in vertical section. 



In the material of Pycnaspis splendens, new genus, new species, 

 at my disposal, at a rough estimate only 0.4 percent of the specimens 

 (6 out of a total 1,500, or more) display both the immature and mature 

 types of ornamentation simultaneously (e. g., the holotype, pi. 1, 

 figs. 1,2). If such specimens were lacking altogether — which might 

 easily have happened in a less comprehensive material — one should 

 no doubt refer the plates and scales, where the ornamentation is of 

 the immature type (fig. 2,a), to another species and (in view of the 

 microstructure of the tubercles, see below) probably even to another 

 genus than those where it is of the mature kind only (fig. 2,6). 

 It is true that, as far as one can tell, a correspondingly marked 

 and abrupt change in the ornamentation of the exoskeleton with 

 advancing age is not met with in any other representative of the 

 Heterostraci and rarely, if ever, in placoderms and early teleostomians 

 either (cf., e. g., in Bothriolepis, Stensio, 1948, pp. 169-171, 376-386). 

 But the condition of Pycnaspis splendens, new genus, new species, 

 referred to above nevertheless shows that in dealing with early Paleo- 

 zoic bone bed material from the taxonomic point of view one has, at 

 least in certain cases, to allow for a considerable margin of error. 



It is worthy of mention, finally, that the ornamentation of the plates 

 and scales was subject to variation not only in different stages of 

 growth but even to some extent in different parts of the body. Thus 

 in the polygonal plates, which presumably lay mainly on the anterior 

 portion of the carapace, aU tubercles, large and small alike, rise straight 



