386 Dr. J. E. Duerden on the 



Similar results are given by Bourne (1887) for the genera 

 Euphyllia and Mussa, and by Fowler (1888) for Lophohelia, 

 all of which reproduce by fission *. These authors state that 

 directive mesenteries are absent in the polyps examined by 

 them and that the septa are only irregularly arranged in 

 orders, and are not necessarily a multiple of six. 



The special characteristics of the polypal system of fission- 

 corals thus summarized can be fully understood only from a 

 knowledge of the early stages of growth. For this purpose 

 the phenomenon of polypal fission as it occurs in Manicina 

 areolata and Favia frayum will be briefly described. 



The development of the larval polyp of Manicina has been 

 followed as far as the establishment of the twelve primary 

 mesenteries, the first two cycles of tentacles, and the appear- 

 ance of the six primary septa. Adherent to the dead parts of 

 old colonies are frequently found small independent polyps, 

 evidently derived from larvse which had fixed themselves on 

 the parent colony. From these all the subsequent stages of 

 development, showing two, four, or more oral apertures, have 

 been secured. Thus for purposes of the present study the 

 development of Manicina is known from the earliest appear- 

 ance of the mesenteries until fission has become fully 

 established. 



The primary tentacles, mesenteries, and septa of Manicina 

 are arranged in regular hexameral cycles. The mesen- 

 teries throughout arise in a bilateral order according to a 

 definite sequence, but in the establishment of any cycle the 

 members become equal, as if they had arisen simultaneously. 

 The furthest stage obtained in the growth of a simple 

 unioral polyp is diagram matically represented in fig. 1. 

 ■Twelve pairs of mesenteries (I, 11), including two pairs of 

 directives (D, D), represent the first and second orders, and 

 extend as far as the stomoda3um ; alternating with these are 

 twelve other pairs of mesenteries, forming a second cycle 

 and representing the third order (111) ; alternating with 

 both cycles is a third cycle of twenty-four pairs (IV). 

 The larval polyp is thus built upon the normal hexameral 



* Lacaze-Dutliiers (1897, p. 154) shows that in Lophohelia proUferiim 

 a portion of an older calice takes part in the formation of a bud (blasto- 

 zijite). As here uuderstood, this would be regarded as an instance of 

 fission in which the two parts become almost wholl}' separated, and 

 Lacaze-Uuthiers apparently regarded it as such. In connexion with 

 the differences between gemmiferous and fissiparous corals, in species 

 where the polyps become almost wholly distinct from one another, the 

 hexameral cyclical regularity of the septal plan in the gemmiferous 

 Amjiltihtlia and its irregularity in the fissiparous Lophohelia, as, shown on 

 pi. viii. ot Lacaze-Duthiers' paper, are veiy instructive. 



