ON LIGHT. 395 



about a common centre, and intersected, not by the two 

 arms of a black cross as in Iceland spar, but by one 

 vertical dark bar cutting centrally across them. This 

 dark bar is converted to a white one, and the colours 

 of all the rings changed to their complementary ones, 

 by turning the analyzing plate through 90 in its plane. 



(170.) In mica, the angular separation of the optic 

 axes is too large to allow both these sets of rings to be 

 seen at once, so as to examine the nature of their 

 mutual connexion. In nitre however, in which it is 

 only about 5 (within the crystal), this may be very con- 

 veniently done, by cutting from the clear, transparent 

 portion of a large hexangular-prismatic crystal (such as 

 may always be found in searching over a lot of the 

 ordinary commercial saltpetre) a plate about a quarter 

 of an inch thick, perpendicular to the axis of the prism, 

 and polishing its faces. If this be placed between two 

 crossed tourmalines, and held up against the light, the 

 normal phaenomenon of the biaxal rings will be seen in 

 its utmost perfection, as in fig. 17, the upright and hori- 

 zontal lines in which indicate broad brushes as it were of 

 shadow, cutting across the system of ovals, and breaking 

 them up into four similar quadrants. If, retaining the 

 tourmalines in the same position, the nitre plate be 

 turned round in its own plane, this cross breaks up 

 into two curved arcs, as represented in fig. 18, cor- 

 responding to a movement through a quarter of a right 

 angle, then, as in fig. 19, corresponding to 45 ot 

 change, and so on till after a quarter of a revolution 

 the original appearance of fig. 17 is restored. If the 



