THE ACTION OF TWO SUPERPOSED CRYSTALLOID BODIES. 345 



The gradations of colour do not of course even here agree with 

 those of Newton's Table, though the rise and fall on increase of e 

 are, as a rule, distinctly expressed. The highest acceleration and 

 the lowest retardation colours are evidently given by the position 

 = 0; if e becomes equal to 90, we get in all positions the red of 

 the selenite plate. Where the change of colour is exceptionally 

 indistinct, it may be rendered more perceptible by the application 

 of another selenite plate, e.g., grey I. 



If we turn the pair of plates, without altering the angle e, upon 

 the selenite plate round a perpendicular axis for instance, by gra- 

 dually moving them from the position of acceleration to that of 

 .retardation we observe that the changes of colour do not agree 



either with Newton's Table or with the colours of a single plate in. 

 the corresponding positions. While a single plate, by rotation 

 upon a selenite plate, produces tints which may be obtained on the 

 palette by mixture of the acceleration and the retardation colours 

 with that of the selenite plate, the pair of plates produce a change 

 of colour of complicated nature, which can be definitely formu- 

 lated only for the more simple cases. The practical significance 

 which attaches to this change nevertheless demands a detailed 

 explanation. 



The difference of colour in the two orthogonal positions, where 

 the bi-sectional line of the angle e coincides with the polarising 

 plane of the lower or upper Nicol, may be denoted as the chief 

 peculiarity of the changes which are observed on rotation of the 

 pair of plates upon a selenite plate. In fact, these positions 

 cannot, under any circumstances, be exactly equivalent, as may 



