ASTEROIDEA OF NORTH PACIFIC AND ADJACENT WATERS — FISHER 111 



sists in the adult only on the superomarginals. This is its history in other genera, 

 as, Astrometis. 



The loss of spines is apparently due to absorption, the material being trans- 

 ferred to new plates and to persisting spines which show a relative increase in size. 

 Scars are sometimes left on plates to mark the place where spinelets stood at an 

 earlier stage. 



In S. heteropaes 14 out of 15 young are six-rayed. Twelve out of these fourteen 

 have four madreporites symmetrically placed with two on either side of the plane 

 of fission (the two opposite interradii through which the disk splits being without 

 them). Thus each half, after fission, has two madreporites — one on either side of the 

 central ray of the triad. (Text figs. 1-3.) The exceptions are a tiny symmetrical 

 specimen, with R 7.5 mm., on which I can find only one madreporite; and a speci- 

 men with two pores on the regenerating half and only one on the original half. 

 (Fig. 3.) At this stage they are normally very small and hard to find. The material 

 of S. euplecta is not so favorable for a census since on regenerating specimens the new 

 rays are small, but in the better preserved sLx-rayed specimens there are three or 

 four madreporites placed as in heteropaes. 



Of the nine Jive-rayed specimens of S. euplecta, eight have one madreporite 

 and one has three madreporites (station 3859). These three madreporites are veiy 

 small and are smaller than in the other specimens. The only five-rayed young of 

 S. heteropaes has two madreporites in neighboring interradii. 



It is clear that active fissiparity is correlated with six rays and with usually 

 four symmetrically placed madreporites; for none of the five-rayed examples shows 

 evidence of having split through the disk. In these only separate rays have been 

 shed as in ordinary autotomy, the disk remaining entire with the five oral angles 

 uninjured. In fissiparity two opposite oral angles are split neatly in twain. 



The location of the madreporites with reference of the plane of splitting would 

 provide two directly opposed "physiologically anterior" points (Cole) and would 

 thus automatically favor an equal splitting of the disk. Crozier 45 regards the mul- 

 tiplication of madreporites at separated points on the disk of Coscinasterias tenui-spma 

 as furnishing an assurance that portions of the body separated by autotomy will 

 each be provided with a madreporic canal. This seems reasonable. However, a 

 large, nonfissiparous species, Acanthaster planci, with upward of 16 rays has 4 to 8 

 madreporites. 



The utility of several madreporites in fissiparous species would appear to bo 

 clear. But as to origin it is not evident in Sclerasterias that the extra madreporites 

 are solely post-larval developments as a preparation for fission. Furthermore, we 

 have a transitory post-larval hexamerous symmetry to account for in a character- 

 istically pentamerous genus. The six-rayed young with four madreporites may 

 have descended from larvae with four hydropores. If so it is likely that wo 

 have in nature the sort of hydropore duplication reported by Newman 4S in labora- 

 tory cultures of Patiria miniata. This physiological twinning in the larva may he 

 here a normal precursor to a subsequent post-larval "untwinning", by which the 

 six-rayed, four-pored, fissiparous young becomes a five-rayed, aonfissiparous adult 

 with one madreporite. [I have specimens showing this last stage, before the spines 



« Biol. Bulletin, vol. 39, 1920, pp. 122-128. 



<« Idem, vol. 40, pp. 118-125. Joum. Eiper. Zool., vol. 



33, pp. 321-352. 



