J3^ 



ambitus, bul then ceasing more or less, as for instance Codiopsis, Gymnodiadema, 

 Plistophyma and Polycyphus; and also Orthopsis and its allies) be classified by the 

 ornamentation of the tubercles, groups possessing very diverse ambulacra wil be asso- 

 ciated". This mayjperhaps be possible; but it is certain that by leaving this feature 

 out of consideration Duncan has been induced to make e. g. Hemipedina a subgenus of 

 Diadema, two forms so different that they must certainly be referred to two different 

 families. And nobody will deny, I am sure, that the recent forms must afford the 

 test of the value of characters used in the classification of the fossil forms. — 

 By using the perforation of the tubercles as a systematic character among the 

 recent Ectobranchiata no forms are grouped together, which by their other charac- 

 ters are shown to stand apart from each other. This is then a character of great 

 value. Among the Diadematids no instance is known, of allied genera having the 

 one perforated, the other imperforated tubercles. Among the Cidarids there is one 

 instance: the genus Tylocidaris Pomel having imperforate tubercles, whereas in all 

 the other recent and fossil Cidarids the primary tubercles are perforate. The crenu- 

 lation of the tubercles is a less reliable character, as there are instances among the 

 recent forms where undoubtedly allied genera, such as Echinothrix and Lissodiadema, 

 Temnopleurus and Pleurechinus, have the one crenulated, the other noncrenulated 

 tubercles. It is, however, in many instances a very useful character, which ought 

 not to be neglected. Further the structure of the spines is of some importance; by 

 neglecting the characters afforded thereby Duncan is induced for instance to make 

 the important fossil genus Pseudodiadema, characterized by its smooth, solid spines, 

 synonymous with Diadema, whose spines are hollow and verticillate. As with the 

 crenulation of the tubercles, however, the characters afforded by the structure of 

 the spines must be used very cautiously. — Finally the pedicellariæ and spicules 

 afford some characters of importance, though not so many as in the other regular 

 Ecliinids. That the characters afforded by the structure of test ought not to be 

 neglected either, needs scarcely be mentioned. — Having thus made clear which cha- 

 racters can be used in the classification we may proceed to discuss the relations 

 of the different genera. 



The genera Plesiodiadema and Aspidodiadema are undoubtedly the most 

 primitive of recent Diadematids. Duncan establishes for them a special family, 

 Aspidodiadernatidœ, distinguished from the Diadematidce by the large, narrow, ringed 

 apical system, formed by broad basais and broad intervening radial plates, by having 

 few interradial plates, each with a large primary perforate and crenulate tubercle, 

 and by the straight ambulacra with numerous low primary plates, with or without 

 primary tubercles. Though I find none of these characters very important or exclu- 

 sive, I quite agree with Duncan that these genera ought to form a separate family. 

 A few characters are to be added from the spicules and pedicellariæ. The spicules 

 are elongate, mesially widened and fenestrated plates, whereas in the Diade- 

 matidce proper the spicules are triradiate or larger irregular plates. With regard 



6* 



