CLARK: THE STARFISHES OF THE GENUS IIELIASTER. 63 



side ; and in another specimen of the same species there are 25 normal 

 rays and ten young ones side by side. 



We are now in position to answer the questions raised by Ritter and 

 Crocker (1900) concerning ray multiplication in Labidiaster and to com- 

 pare the process in that genus and Heliaster with what takes place in 

 Pycnopodia. The questions may be taken up in the order in which they 

 were asked. 



(1.) Do the new rays come in in distinct generations? They do not, 

 but develop entirely independently of each other. A considerable num- 

 ber may develop at approximately the same time, often as many as six 

 or seven and sometimes eight or nine in H. polybrachius, but they show 

 no definite relation to each other. 



(2). Do the successive rays arise at the same and definite places ? 

 There is much evidence to show that they tend to arise in all four quad- 

 rants of the circumference of the starfish about equally, but successively 

 rather than simultaneously. This order is by no means consistently 

 adhered to, however. 



(3). With what number of rays does adult life begin? In Heliaster 

 there can be little question that the number is five. There is no evi- 

 dence yet known in the case of Labidiaster. 



(4) . Are the new rays disposed bilaterally ? Not as a rule ; this point 

 is discussed more fully below. 



(5). Is there a ray corresponding to ray A of Pycnopodia ? Appar- 

 ently not. 



The symmetry of Heliaster, referred to under question four, requires 

 a few words of description. Perfect radial symmetry is of course out of 

 the question, as there is only one stone-canal and madreporite, but leav- 

 ing those organs and the racemose and rectal glands out of account, ap- 

 proximate radial symmetry is possible in Heliaster, apparently only in 

 the 5 -rayed stage ; for the interradius, 5-1 rarely develops any accessory 

 rays and never as many as the other interradii. Bilateral symmetry, 

 however, if we except the racemose and rectal glands, is clearly shown 

 by some individuals, but the plane of division is quite different from that 

 which Ritter and Crocker (1900) show is the adult plane in Pycnopodia. 

 For while in Pycnopodia, the madreporite lies always in the second inter- 

 radius to the left of the posterior half of the line of division, in Heliaster 

 the only possible plane of symmetry is through the madreporite. In 

 Pycnopodia moreover the plane is determined by the position of the 

 accessory rays and every normal individual is bilaterally symmetrical (ap- 

 proximately of course), while in Heliaster the accessory rays have no 



