HELIOLITIDiE. 337 



are lamellar, and extend to a considerable distance into the interior 

 of the visceral chamber, sometimes meeting to form a reticulated 

 pseudocolumella (H intricatus\ or even showing alternate large 

 and small septa (as in some examples of Heliolites porosus). Lastly, 

 the septa may be in the form of twelve longitudinal rows of spines 

 (as in Heliolites dubius, fig. 217, a and b, and in the true Heliolites 

 megastoma of M'Coy). 



As regards the mode of growth of the corallum in Heliolites and 

 its allies, the siphonopores are usually developed fissiparously (fig. 

 217, d), though intermural gemmation also occurs. On the other 

 hand, the autopores are produced by " ccenenchymal gemmation," 

 in a fashion essentially similar to that which has been already de- 

 scribed as occurring in Heliopora. In Heliolites diibins, in which the 

 siphonopores are reduced to a minimum (fig. 217, a and b), a single 

 siphonopore may sometimes be observed to be developed vertically 

 and directly into an autopore. More usually, an autopore is pro- 

 duced by the arrested development of a group of siphonopores 

 (as in Heliopora) ; and the process can be observed readily both in 

 transverse and longitudinal sections. In the former (fig. 217, c) 

 the enclosure of a group of siphonopores within an external wall, 

 and the gradual development of the septa can be readily made out ; 

 while in long sections a few of the siphonopores are seen to be 

 suddenly arrested in their growth, and commonly to be cut off by 

 a common tabula, their place vertically being taken by a single 

 autopore. 



As regards their zoological position, the Heliolitidce show points 

 of relationship on the one hand to the Alcyonaria, and on the 

 other hand to the Zoa?itharia. In the general features of the cor- 

 allum (apart from the minute structure of the skeleton) the mem- 

 bers of this family show a striking resemblance to the unquestionably 

 Alcyonarian genus Heliopora, and in both groups we find the sin- 

 gular phenomenon of " ccenenchymal gemmation." This remarkable 

 mode of increase, however, occurs also in the family of the Fistuli- 

 porid(z ; so that undue weight must not be assigned to this alone. 

 On the other hand, in the fact that the number of the septa is 

 almost constantly twelve, while these structures are sometimes alter- 

 nately long and short, and are occasionally in the form of rows of 

 spines, we have a decided approach to various groups of the Zoan- 

 tharia. The small tubes of the corallum of the Heliolitidce have 

 commonly been regarded as ccenenchymal in their nature, but 

 Moseley's observations on Heliopora would strongly support the view 

 here taken, that these structures are really tenanted by rudimentary 

 polypes (" siphonozooids "), and that the corallum is therefore really 

 dimorphic. It has been already pointed out that the fact that the 

 large corallites are formed by an arrested development and appa- 



VOL. I. Y 



