SOLITARY CORALS. 231 



lower part of their courses the spiral nematocysts and the basophile gland cells become 

 less numerous, and the number of eosinophil gland cells is correspondingly increased. 

 The coiled mesenterial filaments were too much macerated to admit of accurate 

 observation. 1 was only able to satisfy myself that they contain numerous large 

 barbed nematocysts, but none of the spiral variety; that the epithelial cells are 

 attenuated and closely packed together in groups, and that there are numerous very 

 large eosinophile gland cells between the groups. 



The endoderm was too much macerated to admit of careful study. The cells are 

 evidently highly vacuolated, and their nuclei are unusually small. Zooxanthellee are 

 abundant in the most external ccenenchymal canals, but are scantily distributed in 

 the deeper parts. In one of my specimens the endoderm was full of granular 

 amcebocytes similar to those described in Heterovyathus. 



3. Dendrophyllia gracilis, M. Edw. and Haime. 

 (Plate IV., figs. 26 to 28.) 



The anatomy and histology of D. ramea has been fully described by von Heider (26), 

 whose observations are so careful and accurate that I have little to add to them. 

 He has given a very full account of the relations of the perisarc (edge-zone or 

 " Randplatte"), and has fully realized the importance of the exosepta and the 

 manner in which the tertiary mesenterial pairs, and the exosepta embraced by them, 

 are formed in the Y-shaped peripheral extremities of these exosepta. The full 

 significance of this observation of course escaped him, as he was unacquainted with 

 Pourtales' account of the development of Balanophyllia, and his memoir was many 

 years anterior to Duerden's recent work. 



Dendrophyllia is an arborescent genus propagating by lateral buds. The soft 

 tissues extend for a considerable distance below the lip of the calicle, forming a well- 

 marked edge-zone or perisarc. The septa and theca are thin and fragile, and there 

 are well-marked external costee corresponding to the septa. Correlated with the 

 presence of costse is the existence of peripheral continuations of the mesenteries in 

 the perisarc, as has been correctly described and figured by von Heider, There is 

 practically no difference between the septal arrangement of D. gracilis and D. ramea. 

 In both there are three cycles of endosepta, and the exosepta alternate with the 

 endosepta, forming an apparent quaternary cycle. Von Heider has given an 

 excellent account of the manner in which the apparent quaternary exosepta meet 

 and unite in front of the short tertiary endosepta, and his diagram (loc. tit., 

 plate xxxi., fig. 7) leaves nothing to be desired in clearness and accuracy. It 

 is evident that the septa and mesenteries are formed in strict accordance with 

 Pourtales' and Duerden's law, and I am able to add this much in confirmation. I 

 made a series of sections through a small lateral bud measuring about 2 millims. in 

 diameter. In this bud there are only twelve pairs of mesenteries. Six pairs, of 



