528 GILBERT 0. BOURNE. 



appearance, due to the arrangement of the fibres, which will be 

 best understood by reference to the figure. Both the surface 

 view and section show that there is a small remnant of proto- 

 plasm with a nucleus in connection with each desmocyte. As a 

 rule the desmocyte is cup-shaped, just as in Heliopora; and iu 

 such a case its centre appears dark when viewed from above. 

 I have examined many hundreds of these desmocytes, freshly 

 removed from the coral without decalcification, to see whether 

 I could find any trace of crystalline structures in them with 

 the aid of the polariscope. I have never found any 

 either in Mussa, Euphyllia, Lophohelia, Madrepora, 

 Astrsea^ Fungia (all of which were spirit specimens), or in 

 the freshly removed tissues of a newly killed Caryophyllia. 

 The striations, whatever they are, are not due to crystals of 

 carbonate of lime. 



It would be tedious if I were to give details of structure and 

 position of these desmocytes in all the forms which I have exa- 

 mined. Fig. 27 shows their relation to the corallum in Mussa; 

 fig. 33 shows their relation to the peripheral parts of 

 the mesenteries in the edge-zone of Euphyllia ; fig. 42 shows 

 one of them in a radial connecting canal of Madrepora. 

 Figs. 36 to 39 show various phases of their development in 

 Caryophyllia Smithii. In the last-named species the 

 calicoblasts, except at the extreme growing edges of the 

 corallum, are small, highly vacuolated, and without definite 

 cell outlines. At a spot where a desmocyte is about to be 

 formed, one, two, or three nuclei become surrounded with a 

 mass of darker, finely granular protoplasm. The next phase 

 is the appearance of a band-shaped or ovoid body in the centre 

 of the granular protoplasm, which already shows faint signs of 

 striatiou, and stains readily with picro-nigrosin or acid fuchsiu, 

 the striae being usually coloured by picro-carmine if this has 

 been used. Usually one nucleus remains in close association 

 with this body ; the others (if more than one combine to form 

 the granular protoplasmic mass) appear to be concerned in the 

 formation of the mesogloeal process which will join the desmo- 

 cyte to the mesogloeal lamina. The striations next become 



