208 BULLETIN 70, UNITED STATES NATIONAL MUSEUM. 



amount of inflation, and corresponding to this the interradial sulcus is more or less 

 evident. There is one four-rayed example. 



Ahaciinal plates. — Two large si)ecimens from stations 3330 and 3331, Bering 

 Sea, in practically the same locality exhibit very nearly the extremes of variation 

 in the size of plates (PI. 37, figs. 1 and 2). That from station 3330, specimen A (fig. 2) 

 has R = 92 mm., while B has R = 91 mm. In A the largest tabula of the median 

 radial region are 3 mm. in diameter, in B 3.5 mm., but in A there are relatively 

 much fewer large tabula and the spaces between them are much wider than in B, 

 as will be seen from the photographs. This difference in the size of the tabula 

 and in the extent of space between is dupHcated in other localities. The larger 

 tabula have larger granules, not a correspondingly greater number; eighteen to 

 twentv-four marginal granules and eighteen to thirty in the central group is the 

 range in large specimens. In dried specimens the marginal granules are concave 

 on the top, the others are flat. Pedicellarise are usually very numerous, occurring 

 on a majority of plates of the papular areas. 



MargiiKil flates. — The greatest variation of all is in respect to the marginal 

 plates. Typically these form a tumid rounded border, but exceptionally are 

 nearly as thin as in leptoceramus. Again, the two specimens mentioned above will 

 serve to iUustrate the extremes of size. In A, which is less typical, the plates 

 are entirely covered with close, flat granules and measure in the interradius 4 mm. 

 by 4 mm., being less regular in size and not at all tumid. The thickness of the 

 edge of the disk is 5 mm., which includes the height of tho combined plates. In B 

 the plates are ver}" regular, 5 mm. long by 6 wide, and the edge of the disk is 8 mm. 

 thick. The superomarginals are strongly tumid, but entirely covered with gran- 

 ules except for a few distal plates. In A there are forty-two to forty-four supero- 

 marginals and about fifty inferomargmals to a side of the pentagon; in B tliirty- 

 six superomarginals and forty-two inferomargiuals. In typical japonicus as 

 described by Sladen the superomarginals all had an extensive tumid bare space on 

 the abactinal surface, and the inferomargiuals a smaller one on the actinal surface. 

 The size of this spot is variable even when it is present. In my Misaki specimen 

 the spots are smaller than those described by Sladen. One of the Washington 

 examples has the bare spaces (on superomarginals only) about as in the Misaki 

 specimen; three others lack them. Out of twenty-four specimens from station 

 3608 ordy three have a bare spot on the superomarginals and inferomarginals. A 

 small specimen from 3331 has fairly large tumid bare spots, while a large example 

 ("B") lacks them. The shape of the plates is variable. In specimens with large 

 marginals the width is greater than length, but when the plates are small and not 

 tumid the proximal plates may be square or even longer than Avide. The extreme 

 in this respect is a specimen from station 4775 which has proximal plates 4 mm. 

 long by 3.25 wide. Japanese specimens vary in this respect, as the type had 

 longer proximal superomarginals than the Misaki example, which has the plates 

 decidedly wider than long throughout. A remarkable variation is exliibited by a 

 specimen from 4774, which has the marginal plates slightly concave, the transverse 

 sutures being on the summit of a tumidity. The number of pedicellarife on the 

 superomarginals varies greatly, two or three are commonly present, sometimes 

 three or four on all except the most distally situated. The inferomarginals some- 



