534 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



cycle of twelve. Later, situated between the exoccelic cycle and the primary entotentacles, the 

 members of the second cycle of entotentacles begin to appear, as outgrowths from the entocceles 



of the second cycle of mesenteries. 



Thus the exotentacles, whether belonging to the primary or .secondary order, appeared 

 before the entotentacles, and from the beginning they constituted the outer cycle, at first with 

 six and later with twelve members. 



Mesenteries. — For about four weeks no increase beyond the six primary pairs of mesenteries 

 took place, and the fifth and sixth pairs remained free from the stomodamm. Then a pair of 

 mesenteries appeared within the dorsal exoccele on each, side of the polyp. Their first indication 

 was as two narrow lines along the column wall toward its aboral termination. These were 

 followed by a pair in the right and left middle exocceles, and later by a pair in each ventral 

 exoccele. For several weeks the pairs remained of different magnitudes, corresponding with the 

 order of their appearance from the dorsal to the ventral aspect (fig. 6, p. 456). After the third 

 month they began to extend across the disk, but, like the fifth and sixth pairs of protocnemes, 

 never reached the stomodseum. 



Corallwn. — Three or four days after fixation, the skeleton was first observed in the form of 

 six radiating septal upgrowths, practically equal in size, and situated within the six primary 

 entocceles, about midway between the outer boundary and middle of the polyp. At the same 

 time a narrow peripheral calcareous ring was formed, its outer surface uncovered by poly pal 

 tissues, and undoubtedly to be regarded as the epitheca. Macerations made later show the ring 

 to be continuous with the iasalplate, which very early made its appearance. A day or two after 

 the formation of the first cycle of entosepta, the six exoccelic members began to appear, in some 

 cases simultaneously, but in others in successive bilateral pairs from the dorsal to the ventral 

 aspects (fig. 12, p. 492). 



During the course of the third week other calcareous deposits took place, some appearing 

 as angulated continuations of the primary septa, and others arising wholly independent. For 

 two or three months the further development consisted mainly in the increase in size and com- 

 plexity of the parts mentioned, the general impression being that of two cycles of septa, a larger 

 and a smaller, having their peripheral extremity enlarged in a Y-shaped manner, but free from 

 the epitheca. The columella was formed partly from independent upgrowths from the basal 

 plate, and partly by centripetal extensions of the entosepta. 



On the establishment of the second cycle of mesenteries, which naturally coiTesponded in 

 position with the primary exosepta, new calcareous formations appeared independently at the 

 periphery of the entocceles, and later fused with the primary exosepta already in the same 

 radius. The peripheral angulations of the primary exosepta became new and independent 

 exosepta. situated within the twelve exocceles. The skeletal chances now going on were some- 

 what obscure and complicated, but according to my interpretation they afford clear evidence 

 that the members of the second cycle of entosepta must be regarded as new formations, even 

 though later they fuse with the remnants of the primary exosepta. The continuations of the 

 primary exosepta remain exosepta. and for the time being constituted the third cycle of septa. 

 A distinct dorso-ventrality was manifest in the development of the septa. 



