330 Dr BREWSTER on certain new Phenomena of Colour 



but by an infinite number ; and when we chip off the smallest 

 fragment, it gives the same colour as the thickest mass. If we 

 have been successful in obtaining an extremely thin edge, we 

 shall find that the brightness of the tint suffers an evident dimi- 

 nution, though the colour itself never changes ; and at the very 

 edge of the splinter, we can descry, with a good microscope, the 

 individual specks from which the colour is reflected. 



We have already seen, that the light transmitted through the 

 coloured spaces in Fig. 1, does not exhibit distinct complementary 

 tints ; and the same indistinctness takes place in the light trans- 

 mitted through extremely thin splinters that give the change- 

 able colours. But when the spar is the 10th or 20th of an inch 

 thick, the transmitted complementary tint is exceedingly dis- 

 tinct, and, by varying the incidence, it changes from yellow, the 

 complement of the blue of the second order, to blue, the comple- 

 ment of the red of the second order. 



Many of the larger cavities, which have a distinct outline, re- 

 flect a white tint, or a mixture of all the prismatic colours, an effect 

 analogous to the white reflections of the Moon-stone, or Feldspath 

 nacree of HAUY. " Some lapidaries," says HAUY, " give the name 

 of Argentine to specimens of this variety whose pearly reflections, 

 in place of proceeding from the interior, emanate from the sur- 

 face, as in pearls *." The effect here described I have examined 

 in a specimen from Norway, but the light certainly proceeds 

 from the interior, though, from the imperfect transparency of 

 the mass, it appears to a careless observer to be produced at or 

 near the surface. In this specimen, the white light is reflected 

 from planes parallel, or nearly so, to one of the cleavage planes ; 

 while in another face of cleavage, we observe an infinite number 

 of small coloured specks of irregular outline pervading the whole 

 of the specimen, but all parallel to one another, and inclined to 



* Traitt, torn. ii. p. 606. 



