492 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



ter and other forms. If it has permanently the biserial arrangement it 

 will form a connecting link with the Pedicellasteridae, which have 

 very few characters, other than the biserial tube feet, that can not 

 be duplicated among the Asteriidae. The type of Hydrasterias^ H. 

 ofhidion Sladen, is figured as having biserial tube feet, however. 



Family PEDICELLASTERIDAE Perrier. 

 Subfamily Labidiasterinae Verrill. ^ 



In Verrill's scheme of classification this subfamily is placed in 

 the Brisingidae, following the usual treatment of Ldbidiaster. I 

 would suggest however, that Lahidiaster has few essential charac- 

 ters in common with the Brisingidae, but rather exhibits greater 

 structural similarity to Coronaster. This genus seems to be more 

 nearly allied to Pedicellaster than to either Heliaster or to any of 

 the recently proposed genera of Asteriidae. I would therefore place 

 Lahidiaster in the Pedicellasteridae. I have dissected a large example 

 of Lahidiaster 7'adiosus Liitken, from the Straits of Magellan. 



Lahidiaster differs from Brisinga, Odinia, Freyella, and similar 

 genera in the following important particulars: (1) Its abactinal 

 skeleton is not duplicated in the Brisingidae; (2) forficiform, or 

 straight, pedicellariae are present; (3) the adambulacral plates are 

 crowded, very short in proportion to width, and entirely unlike in 

 form and armature the same highly peculiar plates of all Brisingidae; 

 (4) the ambulacralia are shorter, especially the dorsal ends, which 

 overlap, or imbricate with, the next adoral ambulacral plate, while 

 in the Brisingidae there is no sign of imbrication, the ambulacralia 

 resembling the centra of chordate vertebrae, with vertical articu- 

 lating adoral and aboral facets. 



In the Brisingidae (in the narrower sense) the abactinal skeleton 

 of the rays is variable, being in the form of transverse, independent, 

 parallel ridges or costae, separated by areas of integument without 

 plates ; or the intervals may be partially or completely filled in with 

 more or less imperfectly developed plates immersed in the body- wall ; 

 or the arches may be absent and a tessellation of thin plates may 

 cover the genital region of the ray ; or there may be thin plates, more 

 or less spiniferous, together with differentiated transverse costae. 



In Lahidiaster the skeleton of the ray is closely similar to that of 

 Coronaster. There is a longitudinal series of trilobate inferomarginal 

 plates, one of quadrilobate or cruciform superomarginal plates, and 

 one of cruciform median radial plates. The marginals and radials form 

 regular transverse series. On the basal portion of the ray there is a more 

 or less irregular zigzag series of trilobate dorsolateral plates. The 

 primary plates either gonnect directly by their slender lobes, or these 



M'erilll 1914a, p. 26. 



