72 THE VOYAGE OF H.M.S. CHALLENGER. 



of the coronal plates is already markedly pistol-shaped, the sutures do not extend hori- 

 zontally, or nearly so from one area to the other as they do in the other Desmosticha, 

 and we can trace a slight bevel in the same direction as that of Astropyga in a section of 

 the junction of adjoining coronal plates. In Astropyga the plates being narrower and 

 more elongated, and, consequently, more numerous, the pistol shape of the coronal plates is 

 very striking, and is as fuUy developed as in some stages of Asthenosoma , lapping slightly 

 along the median line. Another very characteristic feature of the Echinothuridse we also 

 find in Astropyga. I have called attention to the splitting up of the interambulacral 

 plates into irregularly sliaped independent plates, thus producing interambulacral 

 areas which, as in the Pala3echinid8e proper, are composed of more than two vertical 

 rows of plates. The fact that there are more than two vertical rows in the ambulacral 

 areas in the Desmosticha as well as Pateechinidae I have referred to on a former occasion. 

 In Astropyga we find that the large interambulacral plates from the edge of the ambitus 

 nearly to the abactinal system, as far as the external line of primary tubercles extends, 

 are made up of two very distinct plates (PL X.'' fig. 9), so that in Astropyga as well as 

 in Fhormosoma we have an interambulacral area in which the vertical zones are 

 not composed simply of two rows of plates but of four, and in which the primary 

 tubercles of the actinal surface recall very strikingly from their deeply sunken areolae, 

 those of ArchcBOcidaris and Fhormosoma. Astropyga must therefore be considered 

 a genus either belonging to the Echinothuridae, or at any rate possessing some of the 

 most characteristic features of both the Diadematidae and Ecliinothm-idas. The inter- 

 ambulacral plates of the actinal surface of Astropyga difler from those of Phor- 

 mosoma in which the primary tubercles are deeply sunken ; but when seen from 

 the interior, the deep hollow primary tubercles of the genus connect the solid 

 tubercles of Echinothrix and Diadema with the hoUow and deeply sunken tubercles 

 of Phormosoma proper. The shorter and somewhat club-shaped spines of the actinal 

 surface, like those of Mkropyga, resemble somewhat the shorter hollow-tipped spines of 

 the actinal surface of the Echinothuridae. 



In most of the Desmosticha with spherical tests a certain amount of bevelling occurs 

 in the joints of adjoining plates, this bevelling corresponding more or less to the 

 curvature of the test ; and when there are a large number of plates in a thin test, the 

 edges appear parallel, while the direction of the joints is readily traced in genera having 

 few coronal plates and a comparatively thick test. In Echmaraclmius the upper edge 

 of the plates of the ambulacral system within the petals is inclined towards 

 the abactinal system. It certainly seems impracticable to base a classification of the 

 Echinoidea on this character of the imbrication or abutting of the coronal plates in the 

 difi"erent groups of Echinids as has been proposed by Keeping.^ 



Both in the Cidaridae and in the Echinothuridge in which the imbricating plates of 



' Palaeozoic Echini, Quar. Joiu'. Geol. Soc. London, 1876, vol. xxxii. p. 40. 



