40 L. Agassiz on the Echinodermata. 



at the dorsal summit, and that the spines which they bear in 

 this region are less fully developed. If, carrying the examina- 

 tion still further, we remove the spines, we shall then observe 

 that amongst the oviducal and interoviducal plates and the 

 interambulacral plates that bear spines there are some less 

 fully developed, irregular in form, wanting even the mamellae 

 and the spines, and taking their place among the mammellated 

 plates only in proportion as they gradually attain to a larger 

 size. The new plates are at first very small, and may be com- 

 pared to points of ossification which at first grow simulta- 

 neously in all directions, though their lower side completes its 

 formation sooner than the upper, and the upper side is some- 

 times yet incomplete, even when an incipient mammella is ob- 

 servable in the middle of it. In the region of the body where 

 this increase takes place, the membrane which unites all the 

 plates and spreads itself over their surface, forming an articular 

 capsule about the base of the spines, is softer and more spongy 

 than it is in the inferior part where the plates are consolidated 

 and immoveable. It is in fact this spongy mass that deposits 

 the calcareous matter of which the plates are composed : and 

 the spines shoot out in the centre almost in the same manner 

 as the horns of a stag. They do not become moveable until 

 they have attained a certain stage of development, and there 

 is a period in their growth after which their size does not in- 

 crease. Those however which drop off accidentally are replaced 

 by others, formed, as those had been, by the tumefaction of 

 the membrane which covers the plates. We may always ob- 

 serve in a single specimen of the Cidaris all the gradations of 

 increase, from that of the plates which have completed their 

 growth and bear spines several inches long, down to the 

 smallest points of ossification of the plates yet unfurnished 

 with spines. These facts I have ascertained by examining se- 

 veral individual specimens which exhibited all the interme- 

 diate stages of development through which the pieces in ques- 

 tion must pass : and indeed, when we have no direct means of 

 observing the growth of an animal in one individual, the only 

 resource left us is to compare a great number of individuals 

 representing a complete series of all the stages through which 

 the species to which they belong has to pass before their 



