REVISION" OF PALEOZOIC STELLEROIDEA. 121 



however, can be seen in places, on each side of the ambulacral 

 furrows, and alternating with a row of similar but slightly smaller 

 adambulacral pieces, the number of pieces in each side of these being 

 about the same. [In the proximal portion of a ray there are 19 

 margmal plates in 40 mm.]. The marginal pieces seem each to 

 connect with the adjacent range of dorsal pieces, at places in the 

 rays, by a little central salient pomt only, while those of both series 

 are roughened by numerous comparatively coarse granules [all of 

 which bear small spines], and each piece (especially [only] of the 

 marginal row) also shows a small central pit for the articulation of 

 a spine. Tliese spines are larger than those connected with the 

 dorsal pieces, being generally about 0.22 inch in length and 0.05 inch 

 in thickness; they are smooth, straight, rounded, slightly thickened at 

 the attached end, and tapering at the free end to a slightly blunted 

 pomt. [The adambulacral plates also possess these large spmes, 

 which are inserted on the distal inner edge of each plate.]" 



In each interbrachial area there are two pairs of wedge-shaped 

 pieces and a single diamond-shaped interbrachial marginal plate. 

 There are at least 4 (and may be as many as 10) pairs of adam- 

 bulacral plates orally or m the actmal mterbrachial areas, the terminal 

 pair being of the oral armature. 



Ambulacral plates much wider than long, mcrcasing rapidly in 

 width toward the mouth. Each plate is carinated as in P. speci- 

 osus and proximally as in P. magnijicus, with the podial openings 

 confined to the extreme outer edge of the ambulacral furrows. 



Formation and locality.— ¥ound at Cmcmnati, Ohio, in the ]\Iays- 

 ville beds, about 200 feet below the top of the hills. The holotype 

 is in the Dyer collection of the Museum of Comparative Zoology, 

 Harvard University (No. 13). Another specimen from near the top 

 of the Maysvilie at Cincinnati, preserving only the disk and showing 

 best the actinal area, is in the Gurley collection of the University 

 of Chicago (No. 10989). Origmally this individual appears to have 

 been larger than the holotype. 



Remarks. — The relationship of this species is with P. sinnulosus, 

 but the larger size, greater development of both dorsal and ventral 

 spines, absence of columnar arrangement of the abactinal plates, 

 and the presence of five instead of two or three mterbrachial margmal 

 plates wdll readily separate P. dyeri from the latter species. Even 

 though this form appears to have attained a larger growth than 

 P. magnijicus, which it resembles m some characters, it differs m 

 bemg more ponderous in its skeleton and in the far stouter and 

 longer actinal spines. 



Regardmg the generic position of Palseaster dyeri, Mee\<. \vrote: 



"Possibly I should have called it Petraster dyeri, for if the apparent 

 presence of a few disk plates on the ventral side between the mar- 



