IN PLATES, TUBES, AND CYLINDERS OF GLASS. 367 



negative structure from the two external positive structures. 

 The tint at a, or any part of the line be, would be 



T s-; but if the circular plate were part of a plate si- 



.312 B 



rnilar to, and at right angles to ABCD, the tint at a, or any 

 part of the line EOF, would be equal to T ; and as this tint is 

 rectangular to the other tint at a, the resulting tint will be 

 equal to the difference of these tints, or to 



T T -H- - Td °~ 

 1 ~~ X ~~~ .312 B 2 - .312 B 2 ' 



In like manner, it may be shewn, that in every point of the 



Td 2 



circular plate, the tint is represented by — g-«i" » which is the 



experimental expression for it already found. In plates, 

 therefore, that have only a positive structure, the negative 

 structure still exists, but is overpowered by opposite ac- 

 tions. 



We are now prepared to understand how the negative struc- 

 ture re-appears, as shewn in Fig. 11., by giving an elliptical form 

 to the plate. For, the maximum negative tint produced at O, 

 in the direction g h, is no longer counterbalanced by the tint 

 in the direction ef; and therefore the difference of these tints 

 appears at O, with a negative character. As the points e, f re- 

 move from O, or as the ellipticity increases, the tint at O gradual- 

 ly rises till it becomes equal to T, or times the tint at 



g, when the action of the edges at e and f has no longer any 

 influence at O. The very same results are obtained by the 

 conversion of a sphere into a spheroid, and they are explicable 

 upon the same principles. 



The 



