SYSTEMATIC POSITION AND AFFINITIES. 09 



roids. The structure of the skeleton of these, with its " radial pillars " and their 

 interlacing "arms," is certainly very similar to that of Ilydractiiiia. On the other 

 hand, these types differ from Hydractinia in the constantly calcareous constitution 

 of the skeleton, in its massive construction, and in the fact that tho organism was 

 certainly for the most part not of an " encrusting " habit of growth. The resem- 

 blance between the Labechiidce and Hydractinia echinata does not appear to me to 

 be nearly so close as it is in the case of the Actinostromidaa. At the same time, 

 it must be admitted that the general structure of Labechia and its allies is of the 

 " Hydractinioid " type. 



So far as Hydractinia circumvestiens is concerned, there are the special pecu- 

 liarities that well-marked zooidal tubes are present ; that the interlaminar spaces 

 are reduced to rows of irregular chamberlets, and that certain of the radial pillars 

 appear to open upon summit-apertures. Upon the whole, therefore, the Stromato- 

 poroids which most nearly resemble H. circumvestiens are the true Stromatoporai, 

 and not the Actinostromata. 



The genus Stromatopora, Goldfuss, and the forms allied to this are, however, 

 more nearly related to Millepora, Lam., than to the Hydractinia^. This will be 

 evident from the following brief account of the minute structure of the skeleton in 

 Millepora, though on this point I need say little, as the coenosteum of this 

 genus has been fully described by Professor Moseley (' Report on the Scientific 

 Results of the Voyage of H.M.S. " Challenger," vol. ii, 1881). In connection with 

 the present inquiry I have prepared a tolerably large number of thin sections of 

 various species of Millepora for comparison with corresponding sections of the 

 Stromatoporoids, but I have nothing of importance to add to Professor Moseley's 

 description. The skeleton of Millepora, as regards its main mass, is essentially 

 composed of a complex network of anastomosing calcareous fibres, so disposed as to 

 give rise to a correspondingly complex network of anastomosing and tortuous canals 

 (Fig. 9, c c). In the living condition, this canal-system (Plate IV, fig. 5) is filled 

 with anastomosing stolons of the coenosarc. According to Professor Moseley, 

 " the canals form regular branching systems, with main trunks which give off 

 numerous branches, from which arise secondary branches, and from these again 

 smaller ramifications. The whole canal-system is connected together by a freely 

 anastomosing meshwork of smaller vessels, and communicates freely by numerous 

 offsets with the cavities of the pores." 



The general spongy skeleton, constituted as above described, is traversed at 

 intervals by the vertical tubes in which the zooids were contained. These are in 

 two series, differing slightly in size according as they contained " gastrozooids " or 

 " dactylozooids." The " gastropores " and " dactyl opores" may be irregularly 

 distributed, or the dactylopores may be arranged in more or less definite systems 

 round the gastropores. Whatever may be the nature of the zooids contained in 



10 



