216 THE VOYAGE OF H.M.S. CHALLENGER. 



entoderm of these forms, and there is therefore less risk of confusing their derivatives. 

 The cellular constituents of the two layers are different, and in all the species yet examined 

 they behave differently under the influence of staining reagents. The stomodasum also 

 has a different structure from that of the external ectoderm and stains in a characteristic 

 manner. Such a structure and characteristic staining is only found elsewhere in the 

 rounded free extremity of the mesenterial filaments. 



The mesenterial filaments themselves usually form elongate ribbon-like structures 

 which contain a central strand of mesoglcea clothed on each side by entoderm (cf. PI. 

 XV. fig. 5). The main portion of each filament not only has the structure of the 

 entoderm, but also stains in the same manner. The free margin, however, has invariably 

 quite a different structure and stains more deeply, and possesses all the characters of the 

 stomodseal ectoderm. Frequently the mesenterial filaments are branched, as in the genus 

 Leiopathes and other forms. In such cases the distal portion of the mesoglcea is forked 

 and each fork bears a number of entoderm cells on each side for some distance, but its ex- 

 tremity is clothed with a rounded mass of cells, which are indistinguishable from those of 

 the stomoclseal ectoderm. The free extremity of each filament consists of a single lobe in 

 all the species studied, and I have never seen indications of the formation of lateral lobes 

 of different structure, excepting in the rare and doubtful cases of Antipathella 

 subpinnata already referred to. In this connection it is interesting to note that in 

 the genus Cerianthus, the forms studied by von Heider, the lateral lobes (Flimmer- 

 streifen) are much reduced and, as compared with those of many Hexactinise, are quite 

 insignificant. This fact, taken in connection with the absence of the " Flimmerstreifen " 

 in Antipathina?, renders it possible that they may be phylogenetically of later origin than 

 the median lobe. 



In most Antipathinse the ectoderm of the stomoda3um may be distinctly traced to 

 extend for a considerable distance along the free margin of the transverse mesenteries (cf 

 PI. XIII. fig. 5, and PI. XV. fig. 5). In the blastozooids, at any rate, the ectoderm appears 

 to reach such a position before the mesenterial filaments are formed, and consequently 

 any outgrowth of the free border of a mesentery to form a mesenterial filament must 

 push before it a cap of ectoderm cells. Such, at any rate, appears to be the probable origin 

 of the median lobe in Antipatharia, but the point must also be studied ontogenetically, and 

 we already know that, in colonial forms, the development of blastozooids is not necessarily 

 a repetition of the ontogenetic process. 



General Conclusions. 



The leading structural points brought out in this chapter may be summarised in 

 the following manner : — 



The ectoderm possesses the essential Actinian structure, but the nematocysts are 



