OF >fITRATE OP POTASH. 295 



In measuring the dispersive power of this salt, we cannot 

 expect the same accuracy of result on account of the great 

 angle of the prism. Owing to the nebulosity of the first image' 

 it is impossible to measure its dispersive power ; but it evi-' 

 dently corresponds with its low power of refraction. In order 

 to correct the dispersion of the second refraction, it requires a 

 prism of flint glass, with an angle of nearly 60°. With an 

 angle of 66", the dispersion is more than corrected ; but with 

 an angle of 56° the correction is not nearly completed. The 

 dispersive powers due to these different angles, are contained 

 in the foUowincj Table : 



a 



Angles of the flint glass \ ^^ ' tv • \ ^', 



*'. * ■< 60 Dispersive powers, < 0.( 



P"^"^' - 1 56 lo.( 



66", r 0.0613 



1.0573 

 1.0546 



By taking a mean between the two extreme observations, 

 we obtain 0.058 for the approximate dispersive power, — a result 

 which could scarcely have been anticipated from the substan- 

 ces which enter into the composition of nitre. The follow- 

 ing Table will shew the relation which this measure bears to 

 the dispersive powers of other bodies : 



Sulphate of lead, - 0.060 



Nitrate of potash, 2d refraction, 0.058 



FUnt glass, - > 0.048 



Water, - . 0.035 



In order to examine the character of the rays which form 

 the two images, I polarised the light of a candle by reflection 

 from glass, and viewed it through two of the parallel faces of a 

 hexaedral prism of nitre. When the edges or common sec- 



I*P 2 tions 



