62 bulletin: museum of comparative zoology. 



the splitting of the interbi-achial wall begins, it goes on more rapidly, if 

 development is normal, towards the actinal surface, and the interbrachial 

 tissues there soon separate and the pedicels of the new ray appear. The 

 growth of the new ray forces the older rays on either side further and 

 further apart until they are entirely separated, and the accessory ray 

 takes its normal place between them. The growth of the new ray in 

 length is more rapid than its increase in diameter, so that it is relatively 

 more slender than the older rays. In many cases, owing to some ob- 

 stacle, probably an unusually firm calcification of the interbrachial wall, 

 the new ray fails to split that wall actinally and so is forced to grow 

 upward and appear on the abactinal surface. Its subsequent growth 

 may force the walls apart and it then settles down into its proper place 

 and becomes a normal ray. Often, however, the interbrachial wall fails 

 to yield and consequently the new ray is unable to develop, but remains 

 as a rudiment on the abactinal surface, usually near the boundary 

 between the true disc and the bases of the rays. Such rudimentai-y 

 abactinal rays are by no means rare and may attain quite a size, although 

 usually very small. The largest that I have seen is on a specimen of 

 cumingii (U. S. N. M. No. 15523) 170 mm. in diameter; it is 23 mm. 

 long and seven in diameter, with the base about 30 mm. from the centre 

 of the abactinal surface of the disc ; it is also remarkable in that the tip 

 is turned in towards the disc, as though one side had grown very much 

 more than the other. Usually such an abactinal ray is situated between 

 two normal rays, but not very rarely it is directly over a normal ray. 

 Two explanations of this position suggest themselves; the aborted ray 

 may have been forced into its present position by the growth of one of 

 the normal rays, or a later bud has developed a normal ray where the 

 aborted ray failed. — A comparison of the above given description of 

 ray formation in Heliaster with Perrier's (1891) account of the same 

 process in Labidiaster reveals such similarity as to leave no doubt that 

 the process is identical in the two genera. It may be added that 

 Perrier's figures could be duplicated from specimens of Heliaster, were it 

 necessary, excepting only those showing regeneration. Cases of regenera- 

 tion occur in Heliaster, but are not very common. Occasionally the tip 

 of a ray is regenerated after loss, but several specimens show broken and 

 healed rays where no regeneration is visible. Several cases occur of 

 apparent regeneration of a group of rays, as though a large part of one 

 side of the Heliaster had been cut (or bitten) off and the new rays were 

 to replace those so lost ; thus in one specimen of inicrobrachius, there 

 are 24 normal rays and 13 much smaller, obviously young rays, side by 



