of Crystalline Films, 305 



composite quarter-wave plate, with its planes at 45° with the 

 plane of polarization, before a film ground concave to show 

 Newton's rings. Here we have, in an analogous way, in 

 opposite quadrants retarded one of the component vibrations 

 a quarter-wave before entering the selenite, while in the alter- 

 nate quadrants we retard the other component ; and we get 

 similar dislocations. Again, letting the concave selenite come 

 first, and superposing a \\ plate cut in quadrants with 

 their planes alternately horizontal and vertical, we now have 

 the contracting and expanding quadrants, with the perfect 

 circles in two positions, as in the calcite. We may make 

 the demonstration complete by reversing the process, and 

 superposing our last composite JA, plate on the disk of 

 chilled glass*. We now are applying in each quadrant all 

 the retardations equally to either the tangential or radial 

 vibrations ; and hence the rings remain perfectly concen- 

 tric, while they expand or contract as the analyzer is ro- 

 tated: there is no dislocation at all. Finally, either the 

 quadrant or 12-sector ^X plate superposed on this square of 

 chilled glass gives us a very beautiful demonstration that the 

 dislocation of the crystal rings is entirely due to the ^X plate 

 retarding one component ray in the crystal on one side of the 

 plane of polarization or that at right angles to it, and accele- 

 rating the same component on the other side of those planes. 

 Here we have the square perpendicularly adjusted, with the 

 composite plate superposed. When the analyzer is rotated, 

 the reversal of the sectors on the lines of the black cross keeps 

 the figure symmetrical, as in the last experiment. But you 

 observe that the diagonals of the square are covered, each by 

 a single plate or sector; and a mere glance at the screen makes 

 it obvious that if, in this position, these diagonals were covered- 

 as the black cross now is, by the junction-line between two 

 contrary sectors, they would be dislocated, the colours on one 

 side of the line approaching the centre, and those on the other 

 receding when we rotate the analyzer. But we will now bring 

 these diagonals of the square into the planes of polarizer and 

 analyzer crossed, and superpose the sectors again upon the 

 glass, junction-lines now covering the diagonals. You observe 

 that the state of things is exactly reversed ; and the contrary 

 sectors now do keep the figure symmetrical on each side of the 

 diagonals, while, on the other hand, the single \\ plates which 

 now cover the bisecting diameters of the square preserve the 



* In private experiment we can of course do this with a plate of cal- 

 cite; but in a projecting instrument it is rather difficult to ensure the 

 precise axial coincidence of all the arrangements with the axis of the con- 

 vergent light, without which the experiment fails. 



