41 



3IB. P. H, CAEPENTEE OX THE GENTTS ACTLXCttlETEA. 



of extreme variability. In the specimens I have examined it varies from 13 to 39 ; 

 so that, with one remarkable exception (PI. II. fig. 8), the ten primary arms do not, at 

 the most, divide more than twice, while in two specimens with 18 and 13 arms 

 respectively two and seven of the primary arms remain undivided. I believe, how- 

 ever, that, as a general rule, an axillary is developed on each primary arm, and that the 

 amount of further division is variable, but that a tertiary division is probably excep- 

 tional, so that the number of arms in this species will be found rarely to exceed 40. 



It will have been already apparent from the position assigned to Act. polymorpha in 

 the classification given in sect. 20, that I consider the typical number of distichals and 

 palmars in this species to be three, of which the second (d,) bears a long pinnule, while 

 the third or axillary segment (d a) consists of two primitive segments united by a 

 syzygium. A typical specimen of this condition is seen in PI. II. fig. 10. Out of the 

 twelve specimens of this species which I have examined, but four others resemble this 

 one in having all then distichal and palmar series regularly developed. In each of the 

 other seven specimens one or more of the distichal or palmar series is irregular, con- 

 sisting onlv of two segments, the second of which is axillary without a svzvgium. In 

 one very remarkable case, represented in PI. II. fig. 8, one of the palmar series is 

 reduced to a single segment placed on the distichal axillary, being also itself an 

 axillary bearing the brachials directly on one of its articular surfaces, while on the other 

 are two segments which may be called suprapalmar, of which the second [sp.a) is an 

 axillary without a syzygium, and bears two arms. 



Excluding this remarkable case, the comparative frequency of the usual variations in 

 the distichal and palmar series in the twelve specimens of Act. polymorpha examined by 

 me is seen in the accompanying Table. Prom this it appears that out of 111 distichal 



Table I. — Showing the Variations in the Distichal and Palmar Series. 



