44 THE VOYAGE OF H.M.S. CHALLENGER. 



this method, which will be explained further on, but, on the other hand, the mode of 

 formulation suggested by Bell to express the characters of the arm-divisions in the 

 multibrachiate Comatulse left very much to be desired. For the regular forms which 

 have two or three joints in each arm-division and the axillary a syzygy, his notation is 

 probably as short a one as could be devised. But it gives no means of distinguishing one 

 of these types from the other, or from that of Actinometra multiradiata in which both 

 occur together ; and where the successive arm-divisions consist of two joints only, without 

 syzygies in the axillaries, it gives no information at all respecting the number of the arm- 

 divisions, Antedon palmate/, with three axillaries above the radials having the same 

 formula as Antedon macronema with only one. 



Bell's method is totally inapplicable to irregular types like Actinometra multifida, 

 which have syzygies in the distichal axillaries but none in those of the subsequent 

 divisions ; and the consequence is that species with forty arms receive exactly the same 

 formulae (excepting of course for the cirrus-characters) as others with only ten to twenty. 

 I have referred elsewhere l to other difficulties connected with Bell's method of formula- 

 tion, which is neither elastic enough to indicate exactly on what joint the syzygy comes 

 in the distichal or palmar series, nor does it state the number of joints in each division 

 when there are no syzygies in the axillaries. 



For some years before the publication of Bell's suggestions I had been in the habit of 

 employing for my own use a method of formulation which should briefly express the 

 characters of the rays and their subdivisions, and yet at the same time be elastic enough 

 to meet all the variations of Coma £w/a-structure with which I was acquainted, together 

 with any others that I could consider as possible. It was based upon a knowledge of the 

 structure of over two hundred species, which has enabled me to make the following 

 generalisations. 



1. All ten-armed species of Actinometra which have the two outer radials united by 

 syzygy have the first two brachials united in the same way. 



Examples. — Actinometra pectinata (PI. LIII. fig. 15); no Antedon known. 



2. All many-armed species of Actinometra which have the two outer radials united 

 by syzygy, either have (a) all the arm-divisions of two joints also united by syzygy, and 

 the first two brachials similarly united ; or (/3) there may be three distichals of which 

 the first two are articulated and the axillary a syzygy, while the subsequent divisions 

 (if any) consist of but two joints united by syzygy. 



Examples. — (a) Actinometra paucicirra (PI. LIV. figs. 1, 2, 10); (fi) Actinometra 

 typica (PI. LVII. fig. 1). 



3. If the two outer radials are united by bifascial articulation, the two next joints are 

 similarly united, whether there be ten or many arms. In the former case the third 

 brachial is always a syzygy. 



On the Classification of the Comatulas, Proc. Zoul. Soc. Lond., 1882, pp. 731-741. 



