MONOGRAPH OF THE EXISTING CKINOIDS. 123 



original 10 arms, in species with 20 arms autotomy occurs when the original 10 

 arms are not much more than three-quarters of the length of the adult arms, 

 and in species with more than 40 arms much earlier. Thus in the species with 

 comparatively few arms the arms of the adults are often shorter than the arms 

 of the fully grown 10-armed young, and they may even be considerably shorter, 

 both absolutely and in proportion to the other dimensions of the animal. As the 

 arms increase in number the animals progressively decrease in the relative maxi- 

 mum size of the 10-armed young, and hence the arms of these 10-armed young 

 become absolutely shorter than those of the adults, the difference increasing pro- 

 portionately with the increase in the number of the arms. But, relatively, the 

 arms of the 10-armed young are always much longer than those of the adults; 

 and the greater the number of arms in the adults the shorter do they become in 

 reference to the diameter of the disk, and the more do they differ in that relation 

 from the arms of the 10-armed young. 



The casting off of the arms in the 10-armed young of the multibrachiate 

 species does not occur simultaneously on all the rays, nor do the two arms of 

 a single ray drop off at the same time. The original arms are lost one by one, 

 and the adult arms are produced successively as the original arms fall off. Thus 

 a young multibrachiate comatulid passing from the 10-armed to the adult con- 

 dition always exhibits great diversity in the size of the newly formed division 

 series and arms, some being nearly or quite fulh r grown, while others are only 

 just beginning to form. I have not been able to ascertain the order in which 

 the arms are lost, or whether they are cast off in any regular order at all, as 

 material in regenerating stages has not yet been collected in sufficient quantity 

 to enable any opinion to be formed on the subject. 



The method of the reduplication of the arms in the multibrachiate coma- 

 tulids by the peculiar and economically wasteful process of discarding the juvenile 

 arms at the brachial synarthry or at the first brachial syzygy, and subsequently 

 regenerating from the stump an axillary with sometimes additional division series 

 bearing two or more arms, each of which is in every way like the one which they 

 replace, was worked out by the present author as given in detail here. But I 

 afterwards found that Wilhelm Minckert in 1905 had arrived at exactly the same 

 conclusions by a somewhat different process of reasoning. P. H. Carpenter on 

 more than one occasion seems to have felt the possibility of the existence of such 

 a process, but he was never sure enough of his ground to permit himself to express 

 the idea in so many words; and it remained for Minckert, working on the Blake 

 material previously reviewed in part by Carpenter, first to show how the redupli- 

 cation of the arms in the comatulids is brought about. 



Phyloyenetic sequence of the arms on each ray. 



A very conspicuous characteristic of Lamprometra protectus and Pontiometra 

 andersoni (figs. 275-277, p. 213), well illustrated also in various other multibrachiate 

 species, is that the proximal pinnules on the outer side of the outer arms on each 

 postradial series exhibit the features distinctive of the species to a markedly 



