RAYS AND BILATERAL SYMMETRY 253 



limits of individual variation. No tendency to one-sidedness 

 can be determined, as one side is as likely to show excessive 

 growth as the other. Occasionally, however, one or the other 

 member may be suppressed entirely, producing thus the rather 

 rare specimens with odd-numbered rays. 



From what has been said, it will be seen that ray multiplica- 

 tion takes place in this species during a considerable portion, 

 though not the whole, of adult life. The exact time, however, 

 at which any particular pair of rays will appear in any partic- 

 ular individual is subject to a wide range of variation, judging 

 from the difference in size of specimens having the same num- 

 ber of rays. Thus, eight-rayed specimens have been measured 

 in which R = 40 mm. to 80 mm., while ten-rayed ones with 

 R = 35 mm. to 65 mm. may be found. 



TOPOGRAPHICAL AND MORPHOLOGICAL DIFFERENCES BETWEEN 

 PRIMARY AND SECONDARY RAYS. 



The rays of Pycnofodia being, then, of two categories rela- 

 tive to their period of origin, i. e.^ larva,l and post-larval (we 

 assume that at least five, probably all, of the six rays of our 

 youngest specimens are, as in other star-fishes, of larval be- 

 ginning), the question naturally occurs : Are there any struc- 

 tural differences between the rays of the two groups? So far 

 as the rays in themselves are concerned, there are none. 

 When, however, attention is directed to their relation to various 

 structures of the disc, quite unexpected and rather remarkable 

 differences come to light. These pertain to (i) the pouches of 

 the cardiac stomach ; (2) the radial apical muscle bands ; and 

 (3) the racemose glands. The gastric pouches turned toward 

 the primary rays are larger and more clearly set off than are 

 those turned toward the secondary ones. The muscle bands of 

 the secondary rays show the history of their origin in their ar- 

 rangement in an interesting way, to be described a little later. 

 The racemose glands are never more than ten^ sometimes only 

 nine in number^ whatever the nimiber of rays in the star ; and 

 they clearly are related to the primary rays, new ones not be- 

 ing produced for the secondary rays (fig. 7). Their exact ar- 

 rangement we shall speak of later, but may state here that such 



