48 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. 



is the one least often resorted to. The accessory pieces appear first 

 chronogenetically on the disk, separating the primordial ossicles 

 one from another in various ways, and as the body cavity grows 

 larger in the course of geologic time, they wedge all of the primary 

 columns apart proximally except the ambulacrals, which compensate 

 for the widening of the rays by growing in size transversely. Some 

 accessory plates also develop in the interbrachial areas separating 

 the interbrachial marginals, as in Trimeraster and Xenaster, while 

 in the large-disked forms, as Petraster, their number becomes excessive. 



Ocular plates. — The large ocular plates seen in so many living 

 Phanerozonia are very rarely developed in Paleozoic forms. In 

 fact, the writer knows of but a single occurrence, in the Lower Car- 

 boniferous genus Neopalxaster. 



It seems to the author that these ossicles should have considerable 

 significance in phylogeny because their large growth indicates that 

 they are much older plates than any of the small adjacent pieces. 

 In other words, in all genera where the ocular or sensory tentacle 

 does not lie in an enlarged and grooved plate, the newly developed 

 ossicles of either the ventral or dorsal side appear distaUy to the pre- 

 viously borne pieces, i. e., at the tips of the rays, whereas in the stocks 

 having ocular plates the new skeletal arrivals appear not at the tips 

 of the rays but on the proximal side or inside of the eye ossicles 

 Therefore, where ocular plates are present, these pieces are not only 

 enlarged in size with age, but progressively pushed farther "and farther 

 outward as well by the growing ray tips that lie on their inner sides. 

 This mode of growth may have originated once or several times, 

 there being in the latter case parallel developments of ocular plates. 

 This is a study that can not be pursued far among Paleozoic genera, 

 but in the later forms may be productive in the discerning of 

 phylogenetic lines. 



After the above was written the writer read in Verrill (1914: 20-22) 

 that Fewkes has shown the ocular plates to be the first to appear in 

 the young, and that with growth they are pushed distally because 

 of 'Hhe budding in of new plates between the apical plate and the 

 one next to it." These observations on living starfishes are therefore 

 in harmony with the study of Neopalseaster, and go to show that the 

 bulk of Paleozoic starfishes insert the plates of the primary columns 

 in a different way, i. e., always distally. On this ground alone most 

 of the Paleozoic starfishes should be grouped by themselves and 

 apart from most of the subsequent Phanerozonia. 



Centro-dorsal.— In the center of the disk of many Paleozoic genera 

 of Asteroidea and of some Auluroidea is seen a more or less prominent 

 plate, which is thought to be the equivalent of the centro-dorsal of 

 the embryo of the living crinid Antedon (Comatula). It is a promi- 

 nent though small plate in most of the forms of Hudsonaster and in 



