Division m the Echinodermata. 331 



in which one group of arms is less developed present all con- 

 ditions of size and all possible degrees of development, we 

 must conclude either that the division, supposing it only to occur 

 once, may take place at very different periods in the life of the 

 animals, or else that it is several times repeated, certainly 

 more frequently than in A. tenuispina, at least four or five 

 times, perhaps much more frequently. The comparative rarity 

 of the regular individuals with 6 equally developed arms 

 seems to me to furnish a decisive proof of the frequent repeti- 

 tion of the division ; but it is only by observations upon the 

 living animals preserved for a long time in aquaria that this 

 question can be settled. It may, perhaps, be supposed that 

 the division ceases when, by the slow growth which accom- 



{)anies it, the individuals have attained the limit of their deve- 

 opment, and acquired the faculty of reproducing in the ordi- 

 nary manner ; but at present this is only an hypothesis. The 

 largest of my specimens (radius 46 millims.) is at any rate far 

 from having attained the point at which all trace of division 

 has disappeared ; it has 7 arms, 4 of which (regenerated) are 

 still only 15-28 millims. in length. 



If we refuse to admit that these abnormal phenomena, con- 

 fined, however, within definite limits, may be explained by 

 repeated division and regeneration of the deficient half, it 

 would be necessary to suppose that what occurs is simply an 

 elimination of a certain number of arms efiected for some de- 

 terminate purpose ; and the first explanation which would pre- 

 sent itself to the mind would be that abeady indicated by 

 M. Steenstrup — namely, that in this case we have something 

 analogous to the formation of the hectocotyli in the Cephalo- 

 poda, a group of arms charged with semen being thrown off" 

 to fulfil a special function of reproduction and afterwards rege- 

 nerated. But there is absolutely nothing in favour of this 

 hypothesis ; and it seems to me to have against it the fact that 

 the regeneration of the arms is quite as frequent in the smallest 

 as in the largest specimens. The analogy with the Ophiurida 

 with a supposed spontaneous division, in which this hypothesis 

 would meet with still greater difiiculties, is also a powerful 

 motive for rejecting it for the Asterida also*. 



* An investigation of the organs of reproduction could not elucidate 

 the phenomena under consideration, unless it be made upon a great num- 

 ber of fresh specimens of all ages and all degrees of development. What 

 is especially necessaiy to be ascertained is whether Asterias problema 

 propagates by ova before the spontaneous division is concluded. I have 

 examined some of my specimens with this question in view, but, as might 

 be expected, without throwing much light upon it. I ascertained that 

 the organs of reproduction were well developed in specimens measuring 

 only 14-1 G millims. in radius ; on the other hand 1 found no trace of them 



