538 GILBERT 0. BOURNE. 



deposit, but are seen to be separated from it by a very delicate 

 membrane, which is continued beyond the apex of the cal- 

 careous spine as an unquestionable sheath. Outside of the 

 calicoblasts is the mesogloea, and outside of this again the 

 cndoderm represented diagrammatically. Precisely the same 

 features are shown in all the costal spines of this and several 

 other sections, this particular spine being selected for repre- 

 sentation because the sheath or membrane between the calico- 

 blasts and the calcareous spine is very distinct. There can, 

 I think, be only one interpretation of the phenomena. The 

 spine grows by the addition of minute particles of carbo- 

 nate of lime, which crystallise out of an organic liquid matrix 

 secreted by the calicoblastic layer. These particles attach 

 themselves to the crystalline structure already present, and 

 become oriented conformably with it, so as to continue the 

 pre-existent crystalline figure in the same mode and in the 

 same direction. If anybody doubts the possibility of this, let 

 him study the formation and growth of crystals forming in a 

 concentrated solution of such a substance as potassium sulphate 

 to which a small amount of some colloid substance, such as 

 mucus or gelatine, has been added. A high power of the 

 microscope should be used. As the solution dries upon the 

 slide minute particles make their appearance : at first they do 

 not behave as crystals under polarised light, nor do the 

 particles in the apparently colloidal material lying between 

 the calicoblasts and the corallum of Madreporaria. After a 

 time the particles aggregate themselves at the growing point 

 of a crystal already formed : they increase in number, and then 

 slowly, imperceptibly, they orient themselves in the direction 

 of the axis of the crystal and become a part of it, producing 

 figures exactly like that shown in fig. 43. In the case of 

 the inorganic crystal, the particles, when they are oriented but 

 not yet attached to the crystal, show up with a pale lambent 

 light when viewed through crossed Nicols, the rest of the 

 crystal standing out brilliantly. Exactly the same phenomenon 

 is observed in the growing tip of a costal spine of Madrepora. 

 The conclusion is irresistible, though I only arrived at it 



