SHALLOW-WATER STARFISHES 29 



different styles of skeleton in the more complex genera. (Fig. i, 

 11-iv.) 



The carinal or median dorsal plates (md) usually remain distinct, 

 but are often small and not easily recognizable. The superomar- 

 ginals (sm) are nearly always easily recognizable, though small, and 

 they generally bear spines larger than the dorsals. They never form 

 a stout margin, as in the Paxillosa. 



The inferomarginal plates (im, fig. i) are generally very evident 

 and carry one or more rows of spines usually longer or larger than 

 the dorsals. They are usually separated from the superodorsals by a 

 more or less wide, naked lateral lane or channel on which are rows 

 of papular groups, and frequently many of the larger pedicellariae. 

 In some cases one or more short intermediate rows of lateral plates 

 (intermarginals, m'), bearing spines, may be interpolated in these 

 areas, proximally. 



Both marginal rows normally reach the ocular or apical plate. The 

 peractinal row (fig. i, pa) rarely actually touches the apical 

 plate, and generally falls considerably short of it. 



This affords a useful character for distinguishing the inferomar- 

 ginal from the peractinal plates.^ This name is proposed for the first 

 or primary row of actinal plates. 



To designate collectively the rows of plates that may exist between 

 the inferomarginals and the adambulacrals, I propose to use the term 

 " interactinals" in place of the longer one, "intermediate actinals," 

 used by Sladen and others. I would do away with the use of the 

 words '" actinal spines " or plates used by many writers as including 

 the inferomarginals with the true actinals. Ventrals may be less 

 objectionable if such a collective name be needed, and is so used by 

 Perrier and others. 



The upper marginal row of ossicles is usually pretty clearly 

 defined by special spines. This superomarginal row usually diverges 

 from the lower one, or turns upward at the base of the rays, and 

 may run a little upon the dorsal surface to join its counterpart on 

 the next ray at the dorsal interradial angle or axil. The plates here 

 called subactinals (text-fig. i, so) appear later than the peractinals. 

 They are often lacking in the young, up to 30 to 50 mm. in diameter, 

 even when they are present in the adult. They often extend only on 

 part of the ray, proximally. The number of rows of these in large 

 species increases with the age. 



* It seems desirable to have special and convenient terms to designate the 

 several series of plates and spines in this and related families. 



