52 T. GisLEN, 



(proxiinally and distally on the pinnule) obliterate. The tip of the p. 

 with 2 or 3 smooth segments. Pj + 12 mm. Pg 5,5 mm. P3 and the 

 following ones without a comb. Distal p. 27; 11 mm. The segments: 

 L=1V2 X br, the 4 outermost ones with distal hooks. 



Disk thiown off. The pinnules of some arms without ambulac- 

 ral fiiirows. 



Botli these newly proposed sub-species of the polyform C. parvi- 

 cirra differ in many characteristics from each other. Thus for instance 

 ft has coarser arm-bases and somewhat fewer Br-segments pr cm. but 

 it is chiefly the different appearance of the comb that has caused me 

 to separate these two types and to denote them by different names. 



Within the sub-fam. Coin aster inte of the fam. CumusteiidcL' one can 

 distinguish chiefly H different types of combs. The first of these, which 

 I should like to call the Coinasle)--iy\ie. is characterized by a short, 

 strongly rolled comb, with some few, large and close teeth. The combs 

 occur on every second or third p. to rather far out on the arm. The 

 second, the Comanfhus-iy\^e, has combs with a greater number of teeth, 

 which are small and low, and therefore thinly placed. They become 

 smaller proximally and distally and obliterate towards the tip of the 

 pinnule, leaving the latter smooth. The combs occur on a very few 

 proximal pinnules (never farther than Pg) and in an unbroken series, 

 that is never on every second or third p. as in type 1. The 8'' type 

 — to a certain extent an intei-mediate one — is the Fa/u'«-type, which 

 has I'ather high and large teeth right out to the tip of the pinnule. 

 Combs occur as in type 2 in an unbroken series. 



Both the first-mentioned types are usually very regularly asso- 

 ciated with certain genera within the above-mentioned sub-family 

 Gomastermo. Type 1 is found in the genus Comaster, type 2 is, as far 

 as I have been able to verify, characteristic of the genera Gommitheria, 

 Gomaiithina (?), and Gomanthus (GenoUa). In the genus Gomantlius (Vania) 

 the use of this characteristic, like so very many others, becomes, how- 

 e\er, impossible in the diagnosis. To Gomantlius- Vania are assigned 

 two species distinguished from each other by such an artificial charac- 

 ter as the number of arms. The facts of the matter are vnidoubtedly, 

 howevei', that Gomantlius (Vania) annulata represents one group of forms 



