34 CORALS AND CORAL ISLANDS. 



In other cases, the budding cluster is small, and hence 

 makes small branches, as in the annexed figure of a species of 

 Porites, from the Feejees. The cells in this genus are very 

 small and nearly or quite superficial, as the figure shows. 



New branches are made in such species by a forking of an 

 old one. The budding cluster enlarges as it grows, and, when 

 it is just beginning to pass the regular or normal size for the 

 species, a subdivision of the budding cluster commences at 

 the extremity of the branch. It is a process of spontaneous 

 fission of a branch or stem. In this way the forking in the 

 coral of the figure on page 32 was produced, and also the 

 branching in that on page t,-^. 



Sometimes, again, the budding cluster is a linear series; and 

 then a coral with erect, flattened or 

 lamellar branches is made. 



Again, sometimes each branch of 

 the corallum is only the corallet of 

 a single polyp ; and new branches 

 are added by the budding of new 

 polyps from its sides, each to length- 

 en out into a new branchlet. In 

 this manner the coral here figured, 

 and many like it, were grown. It 

 is a common species of the West 

 Indies. 



When the budding is not confined 

 to any particular polyp, or cluster 

 of polyps, but takes place univer- 



CLADOCORA ARBUSCULA. n i i i • i 



sally through the growmg mass, the 

 coral formed is more or less nearly hemispherical ; and often 

 the process goes on with such extreme regularity that these 

 hemispheres are perfectly symmetrical, even when enlarged to 

 a diameter of ten or fifteen feet. A portion of the surface of 

 one of these massive species, called Orbicella cavernosa, from 

 the West Indies, is represented in the annexed figure. In the 

 growth of these hemispheres, the enlargement takes place in 

 the spaces between the polyps ; and whenever these spaces 



