the Disintegrated Surfaces of Crystals. 21 



images of the candle will be developed at 1, 2^ 3 ; and by a little 

 further action of the solvent, these images connect themselves 

 with the central image S, by the radial lines IS, 2S, 3S, inclined 

 110° to each other, and 30° to the principal radiations from S. 

 By continuing the action, other three images start up at 4, 5, 6, 

 but apparently without any radial connexion with S. The 

 principal radiations «S, Z>S, cS begin at this period to grow faint 

 between 4 and 1, 5 and 2, and 6 and 3. Another immersion of 

 the crystal developes the images 7, 8, 9 ; and by continuing the 

 action, the images 1, 2, 3 become the brightest, and the branches 

 A, B, C become more like images at m,n,o. The central image 

 S has now transferred almost all its light to the new images, 

 and another immersion will make it disappear altogether, leaving 

 the central part of the figure perfectly dark, as in fig. 8. 



It is now obvious, that by repeated actions of the solvent we 

 have removed the whole of the original surface of the crystal by 

 which the central image S was formed, and have replaced it by 

 a great number of facets, which reflect, in consequence of their 

 various inclinations, the different portions of the geometrical 

 image shown in fig. 8. If we carry the process of solution 

 further, the figure will undergo successive changes, becoming 

 larger and more discontinuous in its outline. 



The beauty and regular development of these phsenomena 

 depend in some measure on the perfection of the original surface 

 of the crystal, and greatly on the uniform temperature of the 

 water, and the shortness of the period during which the crystal 

 is immersed in it. The successive development of the figure 

 may be pretty well seen upon an artificial surface of the octo- 

 hedron of alum, provided it is nearly parallel to the original 

 surface. When the inclination of the artificial face is consider- 

 able, the optical figure loses its symmetry, and gradually passes 

 into other figures produced by the other faces of the crystal 

 towards which the artificial face is inclined. 



On some occasions I have found the principal radiations united 

 by a beautiful nebulous web of a triangular form (as shown in 



fig- 9)- . . , 



All the figures above described may be seen by reflexion from 

 the smallest portion of the face of the octohedron, and they are 

 often more beautiful on one part than another. The principal 

 radiations are shown in the figures as if they were seen from the 

 centre of the triangular face, in which case they point to the 

 angles j but in all other cases, the radiations are perpendicular 

 to the opposite sides of the triangle. 



If we expose any of the six square faces perpendicular to the 

 three axes of the octohedron to the action of water in the manner 

 already described, and examine the optical figure which it pro- 



