Protandry and Senescence in Corals. 465 



than from the cutting off of masses of the central polyp by its 

 advancing epithaeca. The presumption at present — from the 

 large number of specimens dredged by Semper — is that these 

 small polyps were formed in the latter manner. With knowledge 

 of some of the facts alone — and none of the possible causes — this 

 form must be deemed to represent a third variation. 



The polyps appear to give reliable characters of the species of 

 the genus. In consideration of this and the definite existence 

 of about 3 per cent, of intermediates, the three so-called species 

 must be deemed to be varieties of a single species. This is the 

 first suggested case of discontinuous variability in the Madre- 

 poraria, and it is particularly interesting from the fact that all 

 the three forms were found both in the Maldives and Philippines 

 in each case in the same habitat. According to the rules of 

 nomenclature the species should be termed F. rubrum, the two 

 varieties being var. stokesi and var. nutrix. The names of the 

 type and its varieties exemplify the singularly unfortunate 

 and inelastic character of these rules. The naming of a type- 

 form among the above three forms is obviously undesirable. 

 Further the form, originally described, merely represented an 

 exceptional case of normal variability, probably brought about 

 by environmental or habitative influences, the usual operative 

 cause among corals 1 . 



The largest specimens of F. rubrum are characterised by 

 having 2-i equal septa fusing to one another by trabeculae in 

 the centre of the corallum, forming a false columella. In de- 

 velopment there are at first 6 septa, which subsequently join 

 together. Additional septa are formed and an additional 6 fuse 

 by trabeculae with the first cycle. The corallite at this stage is 

 seldom more than 2 or 3 mm. along the long axis of its calicle. 

 Development now proceeds more slowly, and further septa only 

 commence to fuse in the calicle, when a length of 8 to 10 mm. 

 is attained. The scar of either of the three varieties is usually 

 in the stage with 12 septa fusing, showing additional reason why 

 the size of the scar is not a specific character. 



It is during this pause in the development that the larger 

 mesenteries commence to receive their generative organs. At 

 first there is a mere slight thickening of the endoderm on the 

 two sides of the mesenteries behind the contorted ends of their 

 filaments. This nutritive endoderm is very granular, and cell 

 divisions are not apparent. Small testicular acini appear in the 

 structureless lamella between its two layers. These are at first 

 isolated, but partially by growth and partially by the addition of 



1 Vide "Marine Crustaceans, I. On Varieties," by L. A. Borradaile, F<iuna and 

 Geogr. Maldives and Laccadives, Camb. Univ. Press, vol. I., Pt. n., pp. 193-8 

 (1902). 



