104 Miscellaneous. 



tainly the same ; but to give more solidity to the oral ring, which is 

 formed of very small pieces, it is soldered to a larger piece situated 

 behind it, and on the oral surface of which it forms a projeclion 

 which enables it to be easily recognized. 



Past.ing to the other families we find that the teeth, which are 

 more or less stout and more or less pointed, are never absolutely 

 truncated on the oral side, and that, in the general plan of the 

 mouth, they have acquired a preponderance over the first arabu- 

 lacral pieces, whilst the reverse is the case in the whole family 

 Astcriada^. The odontophore, which is almost simple in the Echi- 

 nasterida), in wliich the teeth are very feeble, appears everywhere 

 else composed of a body and two more or less prominent small 

 apophyses having articular surfaces. These apophyses fit into the 

 cavities which result from the coalescence of the first ambulacral 

 and interambulacral pairs ; aud in this way the teeth, instead of 

 resting upon one plane, are free to oscillate round these apophyses. 

 Peculiar muscles give rise to the movements of separation and 

 approximation of each pair of teeth. 



The following are the new groupings that I propose (the names 

 of the species examined are placed in parentheses) : — 



The genera Echiaister [E. sej^ositus) and CribreUa (C. oculata) 

 belong to the same family (Echinasteridaj), which is clearly dif- 

 ferentiated from the rest. The genus Mithrodia (Ji. davigera), 

 which approaches them, must nevertheless, I think, be se2:)arated 

 from thetn and form the type of a family, Mithrodidae, which also 

 has affinities with the Linckiada?. 



Solaster papposns and A', endeca, contrary to the opinion of Mr. A. 

 Agassiz, are certainly two species of a single genus. Their affinities 

 are with Acanthaster (A. echin'des), with which they must form a 

 family (Solasteridae). The Linckiada?, from which I separate the 

 genus Ffomia {F. milleporeUa), are enriched with the genus Glue- 

 taster {C. long i pes), 'which formed part of the Astropectinidae. A 

 great division must be established in this family : on the one hand 

 the genera OiJiidiaster (0. pyramidatas and 0. o/ihidianus) and 

 Seijtaster (<S. varioJatus) ; on the other the genera Linckia (L. 

 miliaris and L. diplox) and Chcetaster. In the former group the 

 pieces that M. Gaudry called " interambulacraires internes," instead 

 of going from the ambulacral piece to the second row from the fur- 

 row, pass to the third, the pieces of which are larger : we cannot 

 therefore assign them a serial number. 



In the Goniasteridoe there are, no doubt, great divisions to be 

 made ; unfortunately I have only been able to examine too limited 

 a number of types. However, I shall separate Pentagonaster (P. 

 astrologorum), with M'hich I unite Fromia, from the rest of the 

 family, in which I leave the genera Pentaceros (P. reticulatus, muri- 

 catus, and tarritus), Anthenea (A. articulata), Go)t.{odisciis {G. Pleya- 

 dellce), Culcita (C. Schmideliana), and Gymnasteria {G. cariniferd). 

 But, I repeat, the study of other genera will introduce new 

 groupings. 



