GEOLOGY AND PALEONTOLOGY OF THE CANAL ZONE. 383 



They differ in breadth and thickness according to the cycles; those of the last cycle 

 are very thin and often bend toward and join those of the third cycle. The principal 

 septa are exsert, denticulated, and thickened at the wall. The columella is usually 

 well developed and broad. The paliform teeth are distinct, but not very prominent. 

 It sometimes forms hemispherical masses 4 to 5 feet or more in diameter. 



This species appears to be rare at the Bermudas, and probably occurs only on the 

 outermost reefs. The only specimen seen by me from there was from near the North 

 Kocks. (Centennial collection.) It is a hemisphere about 11 inches in diameter, of 

 the typical form. It is common on the Florida reefs and throughout the West Indies, 

 Bahia, Brazil; (YaleMus.);=var. hirta, nov., with elevated corallites; roughly serrate' 

 thin costae and septa; calicles deep, 5-6 mm. broad; septa narrow, perpendicular 

 within, usually 40-44. 



The description of the Jamaican specimen, when taken in connec- 

 tion with the notes by Pourtales and Verrill, gives a good idea of the 

 extent of the variation of the species except in one particular, that of 

 the septal arrangement. The normal, fully developed calioes have four 

 complete cycles of septa; however, sometimes the fourth cycle may 

 not be complete while at others there may be a few quinaries. In 

 the recent specimens the tertiaries usually, but not invariably, 

 extend to the columella. 



The characters common to all of the specimens may be briefty 

 summarized as follows : 



Corallum massive, base epithecate, upper surface flat, irregularly 

 convex, or domed. Calices more or less elevated, diameter from 5 to 

 11 mm., externally eostate, costae normally subequal. Septa nor- 

 mally in four complete cycles, the members of the first three cycles 

 extend to the columella, but the fourth may not be complete, and 

 sometimes there may be a few quinaries. Columellar trabecular, 

 well developed, large, with a papillary upper surface. 



Remarks on the synonomy of 0. cavernosa. — The names 0. roAiata 

 (Ellis and Solander), O. argus (Lamarck), 0. conferta (Milne Edwards 

 and Haime), and 0. cavernosa var. Jiirta Verrill, are definitely placed 

 in the synonymy of 0. cavernosa, and it is thought probable that 

 0. braziliana Verrill, should be referred to it. These names will be 

 discussed seriatim. 



Gregory applies 0. radiata to this species, as he considers the Lin- 

 naean definition of Madrepora cavernosa insufficient, an opinion with 

 which I do not agree. All the Linnaean descriptions are unsatisfactory, 

 but in this instance Linnaeus refers to the figures of Seba, he places the 

 Madrepora astroites of Pallas in its synonymy, and he states "Habitat 

 in O. Americano." Taking all things together, the original diagnosis 

 with the references seem to me sufficient for purposes of identifica- 

 tion — in fact, the brief Latin description is not bad. 0. radiata was 

 supposed to differ from 0. cavernosa by possessing only three cycles 

 of septa. Pourtales states, in the quotation already made from him, 

 that "In worn specimens the last cycle disappears first; for that 

 reason probably Orbicella (Madrepora) radiata Ellis and Solander has 



