STARFISHES OF THE PHILIPPINE SEAS. 513 



numerous, and genital region extended ; first adambulacral plate, and 

 sometimes the first 2, joined to corresponding plate or plates of the 

 adjacent ray, or if rays are less than 10, the joined pair of first mar- 

 ginal plates may be more or less interpolated between the distal ends 

 of the first adambulacral plates. But the first adambulacral plates 

 of adjacent rays are never separated to their proximal ends so that 

 the mouth plates form the apex of the interradial angle as in 

 Brisingella. No matter what the number of rays may be, the first 

 pair of marginal plates is closely apposed for the whole length of 

 the plate, and, as stated above, their lower sides may be squeezed in 

 between the outer ends of the first or interradial pair of adam- 

 bulacral plates. T\nien there are fewer than 10 rays these plates 

 descend and separate the outer ends of the normally united pair of 

 adambulacrals in such a way that the marginals can be seen from 

 below. This seems to be a means of strengthening the actinostomial 

 ring. A nonmuscular symphysis or syzygy between the first and 

 second adambulacral plates, and between the dorsal end of the second 

 and third ambulacral ossicles. Adambulacral plates short, with a 

 crowded, characteristic armature consisting of an aboral, obliquely 

 transverse series of 2 subambulacral spines and 1 or 2 furrow spine- 

 lets, and usually also 1 adoral spinelet. The lateral spines and the 

 major adambulacral spines are typically long and bristling. Mouth 

 plates small, with 1 to 3 suboral spines. 



Rewurks. — This genus is not so sharply differentiated as either 

 Brisingella or Stegnohrisinga, but is readily to be recognized by the 

 crowded adambulacral armature and the curiously modified, capitate 

 proximal subambulacral spines, which are similar to those of Brising- 

 aster de Loriol. 



The species to be included in this group are: Craterohrisinga fa- 

 owpla (Fisher), C. parallela (Koehler), C. cricophora (Sladen), G. 

 alberti (Fisher), C. eucoryne (Fisher), C. analoga (Fisher), G. 

 variispina (Ludwig), G. rtiulticostata (Verrill), and probably also 

 the aberrant G. evermanni (Fisher). In the last species the proximal 

 subambulacral spines are longer and slenderer than is usual in Gra- 

 terohrisinga, and the modified tip is not so heavy as in typical species, 

 nor is the adambulacral armature so crowded. The second sub- 

 ambulacral spine is frequently quite small and easily mistaken for 

 an aboral furrow spinelet, but on certain plates, without any regu- 

 larity that I can discover, this inner subambulacral spine becomes 

 nearly as large as the outer. On such plates there is usually also an 

 aboral furrow spinelet, making 3 in an oblique transverse series, with 

 the lateral spine additional, if such happens to be present. These en- 

 larged, inner subambulacral spines are present on the outer part of 

 the costal region and beyond and are not shown in the figures of the 



