HUDSONASTERID.E. 69 



concluding portion of the last sentence. The pustulose elevations which carry 

 spines are present on supero- and infero-marginalia alike, and the spines 

 themselves are probably present on the supero-marginalia of Coccaster hulbiferiis 

 (see p. 77). 



The family contains the genera Hudsonasfer, Stiirtz ; Siluraster, Jaekel ; 

 Girvanaster, nov. gen. ; Belaster, no v. gen. ; and Coccaster, no v. gen. 



All the American species of this family are included by Schuchert in one 

 genus, Hndsonaster. Slluraster, the only other genus known to him, has not as 

 yet been described in detail, and its generic characters are at present obscure. 

 Schuchert therefore refrains from placing its single species, S. perfectus, in his 

 comprehensive genus Hndsonaster. 



The phylogenetic considerations already advanced teach us that the family, as 

 defined, may contain diverse lineages : 



(1) All the forms in the early stages of the Mesopalseaster, Promopalseaster, 

 and Xenaster lineages would necessarily be included in the family. It is one 

 disadvantage of Schuchert's classification, that it does not separate vertical lines 

 of descent but merely collects together parallel stages of the different lines. 



(2) Other forms probably represent the highest development of offshoot 

 lineages. 



Our knowledge of Palseozoic Starfishes is not sufficiently complete to determine 

 exactly to which of these categories many of the species belong. It is evident, 

 however, that Girvanaster is a highly specialised type and worthy of a new generic 

 name. Belaster and Coccaster also contain species which are quite distinct from 

 the American forms which have as yet been described. It is perhaps significant 

 that the British species of " Eudsoiiaster " are more distinct from the American 

 species than are the more highly developed British and American Mesopalge- 

 asterinaB and Promopaleeasteringe. This may be because the culmination and 

 migration of the forms of ''Hndsonaster" occurred at an earlier period than that 

 of the descendant genera, and the forms have consequently had more time to take 

 on different characters in the two areas. It is also to be noticed that Coccaster is a 

 Silurian genus and so later than any " Hndsonaster " known to Schuchert, who 

 only had forms from the Middle and Upper Ordovician. 



That the known species of "Hndsonaster" are to some extent specialised, has 

 also been noted by Schuchert, who points out that in these forms there are always 

 accessory plates separating the centro-dorsal from the ring of primary radialia 

 and primary interradialia (see Text-fig. 38, p. 76). He notes (85, p. 49) that "in 

 the young of the cryptozonian UrastereUa (U. nlricJd) and in mature Calliasterella, 

 there are none of these accessory pieces present. For this reason it is thought 

 that in the Ordovicic there will be found a small asterid, even more primitive than 

 Hndsonaster, that will be devoid of accessory disc pieces." 



