44 VERRILL 



One of his diplacanthid genera (Podasterias) had P. liitkeni Per. 

 (not Asterias liitkenii Stimpson) as its type, and is, therefore, of 

 special interest in this connection. In a subsequent work * Perrier has 

 added two more genera: Distolasterias, for D. stichantha (Sladen), 

 separated from Stolasterias mainly on account of its two rows of 

 adambulacral spines ; and Sclerasterias, for S, guernei Per., a new 

 and peculiar type. But in this later scheme he has judiciously 

 omitted Diplasterias ( Perrier, non Kcehler, nee Ludwig) . 



After a careful study of a large number of species, belonging to 

 these various divisions, I am convinced that several other genera or 

 subgenera, based on structural characters of quite as much impor- 

 tance, can be distinguished. 



As understood in this work, the genus Asterias (type A. rubens 

 L.) will be used in a sense much more restricted than by Sladen. I 

 propose to keep separate, as genera, most of the groups called sub- 

 genera by Sladen as, for example, Cosmasterias, Smilasterias, 

 Coscinasterias, Hydrasterias and most of the groups more recently 

 proposed by Perrier. 



Some of these generic groups need to be modified in definition or 

 extent, and others should be added. ( See below, under Geographical 

 Distribution, for the characters of antarctic genera.) 



I am well aware that many students of these starfishes do not 

 think it worth while to subdivide Asterias, even to the extent that 

 was done by Sladen and by Perrier. But in the case of so extensive 

 a group, with a world-wide distribution, it will be practically impos- 

 sible to intelligently study or discuss its evolution and distribution 

 unless we recognize the relationship between allied species by means 

 of distinctive group-names for minor groups, recognizable by mor- 

 phological characters. We should aim at a phyllogenetic classifica- 

 tion, and this is more likely to be attained by the recognition of mor- 

 phologically similar groups of species than by an indiscriminate 

 lumping of all these diverse forms in one great genus, containing 

 perhaps two hundred species or more. 



B. SUBDIVISIONS OF THE ASTERIINJE HAVING DEFINITE LON- 

 GITUDINAL ROWS OF DORSAL PLATES AND SPINES. 



Among the numerous species in which the dorsal plates and spines 

 have a constant and characteristic arrangement, there are several 

 divisions, apparently of generic and subgeneric value. Several of 



1 Contr. a 1'etude des Stellerides de TAtlantique Nord, Resultats des Cam- 

 pagnes Scientif., faits par Albert I, Prince de Monaco, fas. xi, pp. 34, 35, 1896. 



