132 Papers from the Department of Marine Biology. 



simpe rods united at short intervals by more or less irregular con- 

 necting bars (fig. 2). From the nodes short spines project toward the 

 periphery. Near the base of the arm these spine-like processes become 

 longer and they themselves give out branches which unite with one 

 another, so that an irregular lattice-work is formed continuous with 

 that in the body of the pluteus. 



In the pluteus possessing 3 larval arms, the transverse rods are 

 shorter and are met by a third rod to which those supporting this arm 

 are attached. All of these bend downward at their ends, giving off 

 long sweeping branches directed posteriorly. In the larva with 5 arms 

 there is also a corresponding number of transverse rods. 



THE PERMANENT SKELETON. 

 PLATES. 



Although the specimens are all of the same age, a considerable varia- 

 tion exists among them as to degree of development of the permanent 

 plates, so that I have been able to obtain a series showing more fully 

 the early development of these and of the simultaneously developing 

 spines than of any of the other structures (figs. 4-16). 



In those specimens showing the most advanced stage of development, 

 the permanent skeleton consists of a circle of 5 plates surrounding the 

 region where the mouth is to form, one of which lies anteriorly, 2 lat- 

 erally, and 2 posteriorly (fig. 16). Outside of these is a circle of 10; 

 2 lie back of the anterior, 1 back of each lateral, and 3 are clustered 

 behind each of the 2 posterior plates. As a consequence the mouth 

 anlage has an eccentric position, somewhat to the anterior of the result- 

 ing skeletal structure. The plates are all similar and are those to 

 which the first spines are attached. Many of the specimens are in this 

 stage and possess 15 spines and plates. But new ones are constantly 

 developing which crowd the surrounding area as well as push in among 

 those of the first two rows, so that later no definite arrangement can 

 be determined. 



Each plate commences as a triradiate spicule (fig. 4), the radii lying 

 in the same plane and at angles of 120 to one another. Afterwards a 

 fourth branch arises from the center and pushes outward, and it is to 

 this that the spine later becomes attached. Figures 4 to 7 and 10 to 

 14 give a series of these developing plates. The 3 primary radii soon 

 divide, each time at about the same angle (figs. 5, 6, 10), so that by 

 repeated division 3 hexagonal spaces are inclosed (figs. 10, 14). By the 

 continued pushing out and rebranching of the radii a plate like that 

 shown in figure 15 is formed. 



SPINES. 



Early in the development of the structure described above, a second 

 spicule, with 6 rays instead of 3, may be observed lying near each plate 

 and external to it (figs. 5 and 6). Each ray of this new structure soon 



