METAZOA COELENTERA. 141 



when a specimen is held to the light. No. 247 is a verti- 

 cal section of this coral that exhibits these characteristic 

 openings by means of which the zoons communicate with 

 one another. This coral assumes different shapes ; it is 

 branching (No. 245, in more natural condition than No. 

 248, which has been bleached), and it is flat and encrust- 

 ing (No. 246, Madrepora convexa Dana). The long septa 

 reach the center, where there is a more or less spongy 

 columella. The walls of the corallites are not distinct 

 from the coenenchyma, which is largely developed ; and 

 the epitheca is wanting. 



Budding takes place from the central and parent form 

 which has remarkable vitality, and also from the sides of 

 other corallites which are larger and longer than most of 

 their neighbors. The skeleton of the central parent zoon 

 is seen cut in two in No. 247. In healthy condition the 

 ends of the branches of Madrepore are always pointed, 

 but in No. 248 there are a number of diseased tips with a 

 puffed out, swollen appearance. 



The corallites in Porites (No. 249, P. claviaria Lam., a 

 branching form; No. 250, vertical section of P. astrae- 

 oides Lam., a rounded form) are crowded thickly together 

 on a level, with no intervening coenenchyma. The septa 

 are imperfect and spinous ; there are no tabulae and the 

 columella is small. The coral is more dense than in 

 Madrepore (compare No. 247 with No. 250), and there 

 is no large parent zoon. 



We have attempted to show that the Coelentera as 

 represented by the Hydrozoa and Anthozoa have a defi- 

 nite form and a body cavity all the parts of which are in 

 communication with one another. Beginning with a primi- 

 tive Palaeozoic ancestor we pass to living free-swimming 

 hydroids that have a direct development, the hydroid 

 growing into a medusa. 



Again, free swimming hydroids become attached and 

 colonies arise by budding. These produce free medusae 



