30 PROF. P. M. DUlirCAN's REVISION OF THE 



and there is a great temptation to consider typical Cidarids as having but 

 a few, say not more than seven, in a vertical row. 



But all the other structures, or the varieties of structures, are to be noticed 

 in combination with few or many plates. The spines of some species 

 are apparently so peculiar that they have been used as classincatory data 

 of primary importance ; and one subgenus has been founded upon them. 

 But A. Agassiz veiy properly now lays down the law that Cidarids cannot 

 be classified by their spines alone. The nature of the apical disk is very 

 diflFerent in some Cretaceous and recent species : in the former it is solid, 

 and with a tall basal ring and a small pentagonal periproct in some well- 

 known forms ; and in the latter it is large, flat, with feebly united plates, 

 the radial plates sometimes separating (with or without the assistance of 

 some periproctal plates) the basals. But there is every link in the inter- 

 mediate chain of structures to be seen, and even in the same species there 

 are variations in relation in the position of the basals and radials, entry 

 or not of the latter taking place within the ring. The size of the miharies 

 and the occasional want of perfection of the tubercles of the interradial 

 plates close to the apex cannot be of any great importance. 



Both Lamarck and Agassiz, and subsequently Desor, insisted upon the 

 subdivision of Cidaris, in reference to the condition of the primary 

 tubercles and the grooved or not nature of the test between the pores of 

 pairs. Desor thus established Rhahdocidaris and Leiocidaris ; Brandt 

 established the subgenus Phyllacanthus from the nature of the sjjines and 

 straightness of the ambulacra. A. Agassiz decides against Leiocidaris, and 

 adds to the diagnosis of Phyllacanthus ; and he introduces Dorocidaris, 

 which is a true Cidaris with non-crenulate tubercles. De Loriol and 

 M. Cotteau have both altered the range of diagnosis in Rhahdocidaris and 

 Leiocidaris; and the first-named naturalist decides against Leiocidaris 

 and Phyllacanthus. 



It appears, after having given the subject careful consideration, that 

 none of these divisions is worthy of a true subgeneric position, but that 

 Rhahdocidaris, Leiocidaris, Dorocidaris, St ephanocidaris, Phyllacanthus 

 (Brandt), A. Ag., and Porocidaris are fairly useful artificial divisions. They 

 are synonyms of Cidaris, however. Goniocidaris is a good subgenus. 



I. Division. Typical Cidaris. 



If the system of subdividing tlie genus artificially is adopted, 

 the typical species will be those wiLh a small number of inter- 

 radial coronal plates (5-8), and with the ambulacra more or 

 less undulating, the pores of pairs rather close and separated 

 by a nodule or ridge, and the primary tubercles perforated and 

 crenulated. 



The other divisions will be as follows : — 



