Brewster's Luminous Figures. 499 



that the incidence of the light may be as nearly normal as possible. 

 The crystal is then turned until the image of the reflexion of 

 light is distinctly seen on the surface, and thereby the eye 

 brought as near it as possible. On the table, and at the place 

 above which the crystal is seen, a dead-black paper is laid. If 

 the transparency is enough to allow transmitted light to be seen, 

 the crystal is held with the thumb and forefinger as previously 

 described, so as to cut off the lateral light as much as possible 

 close to the eye, and the flame looked at through it. It is to be 

 observed that the luminous figure is mostly first distinctly seen 

 when standing at a distance of two or three steps from the flame. 

 In judging the luminous image, it is to be considered whether 

 only one face, or at the same time its parallel, was corroded, 

 because the latter often gives the picture of the first inverted; 

 thus, for example, in one corroded face a three- rayed star is to 

 be seen, but a six-rayed one if the parallel face was also cor- 

 roded. 



The images are very beautiful when the crystalline lamina? 

 are fastened in pieces of cork, and the flame looked at through 

 an opera-glass at a distance of about eight steps, the crystal 

 being held between the eye and the eyepiece. 



Such images are easiest produced and observed in alum. 

 When a moist cloth is passed once or twice over a smooth octa- 

 hedral surface, and then a dry one, a three-rayed star immediately 

 appears in the principal form, like fig. 1, PL II.; by frequent moist- 

 ening, it changes in the centre, and three short rays appear 

 between the first ; but the star is immediately changed into the 

 six-rayed one (fig. 2) when the crystal is moistened with dilute 

 nitric or hydrochloric acid in the manner described. I used 

 mostly one volume of concentrated acid, and one or two volumes 

 of water. Further moistening with water (and drying) alters the 

 six- rayed star again into a three-rayed one. Brewster also states 

 that such a corroded surface, upon which triangles are visible, as 

 in fig. 3, reproduce themselves perfectly if the crystal is dipped 

 in a saturated solution of alum, and that the redintegration and 

 filling up of the attacked parts proceeds in this manner with 

 inconceivable rapidity*. I did not quite find it so, yet I obtained 

 normal surfaces when a corroded strip of alum was dipped in a 

 warm and not too concentrated solution of alum and then left 

 to spontaneous evaporation. The faces of the hexahedron and of 

 the rhombic dodecahedron, which often occur in alum in combina- 

 tion with the octahedron, present this phenomenon — that in the 



" The singular fact in this experiment is the inconceivable rapidity with 

 which the particles in the solution fly into their proper places upon the 

 disintegrated surface, and become a permanent portion of the solid crystal." 

 — hoc. cit. p. 174. 



2 L2 



