170 BRITISH PALEOZOIC ASTEROZOA. 



characters of the adambulacralia. These ossicles are extremely broad, with an 

 oral surface scarcely above the level of the cross-ridge of the ambulacralia, and 

 with a characteristic ornament (Text-figs. 118, 119, p. 174). The alignment of 

 the ambulacralia with the adambulacralia makes the groove very shallow, and the 

 oral surface of the arm flat. This, with the petaloid shape of the arm, produces a 

 leaf-like appearance suggesting the generic name Platanaster for the Ordovician 

 form (irXaruvoQ, a plane tree, so called from its broad, flat leaf). 



Although both Platanaster ordovicus and Palasteriscus devonians have the same 

 peculiar adambulacralia, which, with other characters, show that they belong to 

 one and the same small branch of the Asterozoa, they differ in other ways so con- 

 siderably that one is loth to follow one's first inclination and put them in the same 

 genus, regarding one species as the ancestor of the other. The earlier form has 

 well-defined infero-marginalia which are lacking in the later species. This in itself 

 would not be an insuperable objection to placing the forms on the same exact line 

 of descent, for it is possible that infero-marginalia may be lost during lineage 

 evolution. A far greater difficulty is the position of the madreporite. In the 

 Ordovician species this plate is apical and marginal, while m the Devonian species 

 it is oral and quite near the mouth. Following the line of argument developed 

 later (pp. 178-182), one would place the first form, with its fused arm-bases (Text- 

 fig. 114, p. 172) and its apical madreporite, with the Asteroidea. It is difficult to 

 explain the position of the madreporite in the later form except by the supposition 

 that its arms never fused at their bases, and so allowed a secondary growth of the 

 apical interradii to carry the madreporite on to the oral surface. This would place 

 Palasteriscus devonicus among the Ophiuroidea, and we should have the paradox of 

 two nearly related forms belonging apparently to distinct branches of the Asterozoa. 



Fortunately, we have other series of related forms which show similar 

 phenomena. The arms of Stenaster often approximate at their bases because 

 their adambulacralia are so very broad. This character causes fusion of the arm- 

 bases of Monaster, which, according to Schuchert, is allied to Stenaster. The 

 madreporite remains in consequence apical and marginal in this latter form. On 

 the other hand, Aspidosoma, a near ally of Stenaster, has unfused arm-bases, and 

 a downgrowth of the apical interradii, which carries the madreporite to an oral 

 position. 



We can only suppose that among the earliest Asterozoa the impulse to become 

 definitely either Asteroidea or Ophiuroidea was not fixed. I have suggested below 

 (p. 195) that the "Asteroid" character of fusion of the arm-bases and the 

 " Ophiuroid " character of downward growth of the apical interradii are both 

 brought about by feeding habits. According to my interpretation, Platanaster 

 ordovicus is a form which has specialised in ciliary feeding, and has developed a 

 large ciliary area by increasing the surface of the infero-marginalia, which in con- 

 sequence meet along the interradii, while Palasteriscus devonicus is descended from 



