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



secondary intraseptal, then follow the twelve tentacles of the tertiary intraseptal spaces 

 which are visibly smaller, whilst the twenty-four last tentacles communicate with the 

 interseptal spaces. The difference in the size of the tentacles is, however, rarely so 

 marked as this, for a partial or complete equalisation in their size usually takes place at an 

 early period of development. 



When there are a lar^e number of tentacles there is no room for them in a single 

 row, but they are forced to form several rows, of which the inner are the oldest, the 

 outer the youngest. All the tentacles of the same circle are essentially of the same 

 size, but a difference in size may arise between different circles, which is showm by 

 the innermost, and therefore oldest, tentacles being the largest, the outermost and 

 youngest the smallest. If uniform growth take place in all parts of the body of the 

 Actinia, the whole of the circles are arranged, as may often be seen, in multiples of six. 

 For example, twelve tentacles of equal size form the innermost circle, if it be composed of 

 tentacles of the first and second order ; twenty-four, if those of the third order be added to 

 the number ; the next circle would then be occupied by the twelve or twenty-four suc- 

 ceeding tentacles of the sequence. There are often, however, numerous variations, caused 

 by unequal growth in the different sextants ; for example, the tentacles of the third 

 order may have advanced into the innermost circle in one sextant, whilst in another 

 sextant they remain in the next circle. In this way it may happen that the arrangements 

 of the tentacles and of the septa do not correspond completely, even though they are not 

 directly contradictory, since the tentacles of a later order are, at all events, not laiger 

 than those of a preceding order. I only know one exception to this rule, Polysipkonia 

 tuberosa, to the description of which I refer the reader. 



Hitherto I have only spoken of tentacles which are placed on the margin, and 

 which always remain equivalent to each other even when, changing their position for 

 want of space, they have retreated on to the oral disk, and become apparently arranged 

 in several rows. I have already placed these primary tentacles in ojyposition to the 

 secondary tentacles, which are associated with the primary in the Coralliniorphidaa 

 (PL II. figs 1 and 3). They form a system in themselves, and are placed half-way 

 between the peripheral and oral margins of the disk. They correspond only to the 

 intraseptal spaces, and their size is graduated according to the age of the latter. 

 The above-mentioned rule undergoes an exception here, as each intraseptal space 

 communicates with two tentacles, a primary placed on the margin, and a secondary 

 placed on the disk. The Corallimorpkidse are the only exceptions which I have 

 observed, for the Cerianthidae, which are usually described as Actiniae with a double 

 corona of tentacles, cannot be considered here because of the aberrant arrangement of 

 their septa, which are not united in pairs, thus rendering the distinction between the 

 intraseptal and interseptal spaces impossible. 



The septa bear the reproductive organs, the mesenteric filaments, and in many families 



