SYSTEMATIC CLASSIFICATION OF ECHINI. 213 



Hemicidaris in the upper half of its test has returned to simple ambulacral plates. In my 

 tables of ocular plates (pp. 154-159) Duncan is followed in the grouping of fossil genera in 

 families, and it is possible that some genera placed in this family may have other affinities. 

 The plates are thick and most genera have perforate tubercles. 



The Aspidodiadematidae are essentially characterized by simple ambulacral plates, large 

 primordial ambulacral plates which nearly cover the peristome (text-fig. 51, p. 80), and a very 

 large apical disc, all of which are primitive characters. There are very wide insert oculars 

 (text-fig. 80, p. 104). The test is thin and fragile. 



The Centrechinidae have compound ambulacral plates of usually three elements each. 

 The Mesozoic species have relatively thick plates, not imbricate, oculars usually all exsert. 

 The living species have thin tests more or less imbricate, three or more oculars usually insert 

 (text-figs. 81-99, pp. 105-109); while the sequence of incoming of oculars is typically I, V, IV, 

 II, it is not rarely {Centrechinus setosus) I, V, IV, III, a feature rare in the order as a whole, 

 but a typical variation of many cidarids (p. 96). There are only ten primordial ambulacral 

 plates of moderate size on the peristome with non-ambulacral plates (Loven, 1892, Plate 12, 

 figs. 153, 154), a strong distinction from the Echinothuriidae. While the lantern is typically 

 erect, it is inclined in Astropyga, as in the next family. 



In the Echinothuriidae the test is remarkably thin and strongly imbricating, and ambu- 

 lacral plates are compound, composed of three elements each. The primordial interambulacral 

 plates are retained in the basicoronal row (text-fig. 43, p. 80). It is remarkable that a peri- 

 stome of the size attained in this group comes about by the growth of the plates and not by 

 resorption of the base of the corona. Perhaps the most distinctive character of the family is 

 the fact that the peristome is plated with many rows of ambulacral plates only, a feature rarelj' 

 seen in the Cidaroida and in no other post-Palaeozoic group ; but a character of many Palaeo- 

 zoic genera, as Palaeodiscus (Plate 18, fig. 2), Hyattechinus (Plate 23, fig. 1), and Lepidesthes 

 (Plate 68, fig. 3). Mortensen (1904, p. 53) on the strength of my observations classed the 

 Echinothuriidae with the Lepidocentridae because of the similarity of the peristome. I cannot 

 agree with this view; the ambulacra and interambulacra differ radically from those of the 

 Lepidocentridae. In addition, the auricles, peristomal gills, and pits in the top of the pyramids 

 class this family with the Centrechinoida and with close affinities to the Centrechinidae ex- 

 cepting for the peristome. This is the relationship recently adopted by Messrs. Agassiz and 

 Clark (1909). 



The character of the peristome with many rows of ambulacral plates only may perhaps 

 be best considered a localized reversion, or parallelism to some ancient type. Such is seen, in 

 fact, in the young of all regular Echini where the peristome for a time is covered with ambulacral 

 plates only. The oculars of the echinothuriids are exsert in the young (Plate 3, fig. 8), but in 

 the adult are all insert and often are separated from the genitals (text-fig. 170, p. 149), a 



