SEA. LILIES, STARFISHES, ETC. — CLARK. 83 



a few dots or no markings at all. Diversity is equally 

 marked in other features. The distance of the first arm- 

 fork from the distal end of the radial shields ranges from 

 .25 of the radial shield length to 1.10 ; or the same fact may 

 be expressed in another way by saying there are from 9 to 15 

 arm-joints in the undivided arm-base. The number of arm- 

 divisions ranges from 6 to 10 or more. Most of the specimens 

 show the radial shields well separated, and the disk has 

 distinct interradial areas, but there is a tendency on the part 

 of some individuals to have the radial shields united in pairs 

 by thick, concealing skin and in such cases the interradial 

 areas are much reduced ; thus the condition, characteristic of 

 Conodadus, is approached. Further resemblance to Cono- 

 cladus is brought about in certain specimens by the large size 

 of the disk spines. In typical A. australis, these spines are 

 about a millimeter high, conical, and the thickness at base 

 about equals the height ; they are present chiefly on the ' 

 radial shields and are often confined to them ; in many speci- 

 mens they are smaller and more numerous, and rarely they 

 are reduced to low, inconspicuous tubercles. But in some 

 cases, notably in some South Austrahan specimens, the spines 

 are 1.5 mm. high and correspondingly stout. In these cases, 

 too, the spines on the basal parts of the arms are larger than 

 usual and the resemblance to Conocladus may be quite marked. 

 Nevertheless none of the specimens have the radial shields 

 completely united into five disk-wedges, with very much 

 reduced interradial areas as in Conocladtis, nor do any have 

 spines on disk and arm-bases really as big as in that genus. 

 For these reasons one can, without difficulty, separate speci- 

 mens of the two genera, but I am not sure that the separation 

 is not an arbitrary rather than a natural one. I have previ- 

 ously expressed the beUef that Conocladus is derived from 

 Astroconus, while Doderlein thinks the reverse to be the case. 

 Which of us is correct can only be determined by a study of 

 growth-changes in much younger specimens than any at 

 present available. But the youngest specimens in the present 

 series are not at all Conocladus-Vike, the radial shields being 

 narrow and widely separated, the interradial areas large and 

 the disk spines few and small. It is quite possible that further 

 study and more abundant material will show that Conodadus 

 oxyconus is only an extreme form of Astroconus australis, and 

 not a separate species. In that case the genus Conocladus will 

 need a new name, i.e., the species amhlyconus and microconus, 

 for the name Conocladus will have to replace Astroconus as the 

 generic name of VerriU's Astrophyton australe, since it has the 

 sarhe type and antedates Astroconus by two years. In any 

 case, whether C. oxyconus is specifically different from 



