540 GILBERT C. BOURNE. 



examined Caryophyllia with special care. Fig. 31 is a 

 representation of an optical section of a septal granulation as 

 seen with the oil immersion. The apex of the granulation 

 was, of course, out of focus, and is not represented, but its 

 periphery is seen to be composed of a number of needle-like 

 crystals radiating outwards. Tiie crystals are grouped together 

 in bunches, and all the crystals in a bunch diverge from an 

 ideal centre towards the free surface, not only horizontally, 

 but upwards and downwards in all directions where they can 

 extend. Fig. 32 is a not very satisfactory drawing of the 

 surface of a very thin palus of C3,r yophyllianear its extreme 

 upper edge. The surface is very irregular, and studded with 

 hemisphaeroidal projections, each of which is composed of 

 fibro- crystals, diverging iu all directions where there is room 

 for their growth. This is a much more common appearance, 

 in my experience, than the flattish overlapping '' scales." 

 Similar sphseroids or hemisphseroids are formed in many in- 

 organic solutions, and I opine that we find in these the '^crys- 

 talline sphseroids^' described by von Koch and derided by Mrs. 

 Gordon. It is worth noticing that the extreme irregularity of 

 the surface of the corallum of Caryophyllia corresponds 

 with the incoherent and irregular arrangement of the calico- 

 blasts which I have described for this species. The continued 

 growth and aggregation of such sphseroids would give rise to 

 what mineralogists term a botryoidal formation, and the 

 Madreporarian skeleton may fairly be described as a complex 

 botryoidal growth, whose character has been influenced pro- 

 foundly by the living tissues which gave origin to it. 



If my conclusions are correct the whole skeleton of a Madre- 

 porarian polyp has some sort of analogy to a single Alcyonarian 

 spicule. Both are composed of fibro-crystals having an infinitely 

 various, but for every species a definite and characteristic 

 arrangement. As the spicule is enclosed iu a sheath, and 

 separate from the protoplasm of the cell or coenocyte from 

 which it is developed, so is the Madreporarian corallum covered 

 with a membrane or corallum sheath, and separated from the 

 calicoblastic layer which gave rise to it. If there is no diffi- 



