ASTEROTDEA OF NORTH PACIFIC AND ADJACENT WATERS — FISHER. 21 



the marginals, and tlie presence of supcrambulacral plates, favor the Piianerozonia. 

 The single ampullae are noteworthy. They are double in the families mentioned 

 above. As tlirowing some light on the relationsliip of Ohietaster with the Odontas- 

 teridse, the odd interradial marginal is important, and the presence, in the recently 

 described Chsetaster vestitus Koehler, of incipient recurved speciahzed teeth to each 

 pair of mouth plati>s. These spines have an enlarged opaque base and a hyaline tip." 



With the GoniasteridiB a transition between the paxilliform, or tabulate abactiiud 

 plate, and the flat granulate or smooth type takes place. It would be desirable, 

 if it were possible, to divide the Goniasteridte into two families, one with tabulate 

 more or less paxilliform abactinals, and the other M'ith flat pavement-like plates, 

 either with or without secondary intermediate connecting ossicles of various sizes. 

 The practical difficulties in the way of such a course are discussed under the Goni- 

 asteridfB. The diagnosis of the Goniasterida? as it now stands is rather too general, 

 and perhaps vague. The reason for this is the persistence from past ages of many 

 intermediate intergrading forms. Naturally if there is anything to evolution, these 

 intergrading foniLs should be expected. (If they all persisted, it is evident that 

 there would be no definable orders, families, genera, or species.) The Oreasteridae 

 are not very sharply separated from the Goniasterithr, but after the removal of 

 Amphiaster the group is more homogeneous, and is readily recognized by the stellato- 

 reticulate abactmal skeleton, large papular areas, with numerous small papular 

 pores, and the finely granular skin which hides the underlying plates. The Goni- 

 asteridas, on the other hand, have single papulae, or a ver}- limited number to an area, 

 and abactinal plates tessellate, or in a close network through the presence of small 

 intermediate plates. The transition between the typical tessellate form and the 

 more open structure such as occurs in Hippasferia and Anthenea is too complete 

 to keep these genera, which are not very closely related, in a separate family. They 

 are nearer Goniaster than are Pseudarchaster, Mimaster, or Gephyrcaster. The 

 Asteropidae are also difficult to diagnose satisfactorily, although the component 

 genera differ from the Goniasteridae in having small imbricated marginals, and a 

 variable, loose abactinal skeleton overlaid by thick skin. The Linckiida' (omitting 

 Chsetaster) have come to be more or less traditional. The phanerozoniate charac- 

 ters are rather unstable, especially the papulae. Some of its genera are near the 

 Goniasteridae. 



As alread}^ stated, Ludwig has raised the Benthopectinidae, which he divides 

 into two. families, to ordinal rank. The character upon which greatest emphasis is 

 laid is the presence of a pair of muscle bands extending along the ccelomic surface 

 of the dorsal wall of the ray, from near the base to the tip. These are not present in 

 any other group. The alternate marginals are correlated with the muscles, as greater 

 flexibihty is given to the ray. Ludwig beheves the animals can swim by flex- 

 ion of the raj's, and this idea is incorporated in the name of my new genus, 

 Nearchaster. The pectinate pedicellariaj of several genera are also peculiar to the 

 group. 



In this report I have kejjt the ordinal groups somewhat larger than an accept- 

 ance of the Notomyota would permit, although it is possible the Forcipulata are 



"Koehler, An account of the Shallow-water Asteroidea, Echinodernia oi the Indian Museum, pt. 4, 

 June, 1910, pi. 19, fig. 11. 



