106 PROGRESS IN MICROSCOPY 



to be small, is the path difference between the waves O and E. Both 

 polarizer and analyser (not shown in Fig. 3.12) bring about the inter- 

 ference of the waves. To ensure adequate sideshift, thickness of the 

 plate Lo is to be rather substantial and J wil! not ever be so small 

 as to be able to use white light. In the case shown in Fig. 3.12, the 

 experiment should be performed in monochromatic light. However, 

 the path difference J can be made as small as desired, while retaining 

 a substantial image duplication, by a slight alteration of the device 

 shown in Fig. 3.12. The solution is shown in Lebedeff's layout 

 by the plate L^ and the half-wave Q (Fig. 3.10). This makes 

 work in white light feasible. Every type of interference microscope 

 has its own solution. We will describe in the following para- 

 graphs the chief instruments so devised. For the time being, we 

 are using Fig. 3.12 merely to expound the working principle and 

 it is assumed that any value may be given to J. Neither are 

 we defining more accurately the birefringent layout in relation to 

 the microscope. What is known is that the microscope's objective 

 images the object and in the image duplicated by the plate Lo, the two 

 waves O and E are arranged as shown in Fig. 3.12. Said waves are 

 shown at a greater scale in Fig. 3.13. Both images of the object A 



Fig. 3.13. The duplicated waves. 



are at A'q and A'^. The path difference is 1 in the areas wherein the 

 waves are flat and parallel, i.e. the imageless areas. At A'q and ^4^, 

 the path difference between the waves O and E has not the same value. 

 If () is the path difference originated by the object, the path difference 

 between the waves O and E will be 1 + f") at /i^ ^^"d 1 — f> at A'l^. The 

 light-intensity in the two images A'^ and Ay differs from the intensity 

 in the remainder of the field. The transparent object beconies visible. 

 Let us adjust the instrument so that I equates the 0565 /^ wave- 

 length and let us assume that both polarizer and analyser are crossed. 

 In accordance with the classical data pertaining to polarized-light 

 interference, the first-order purple hue (Newton's scale of colours. 

 Chapter VII, ij 1) is observed in the image-less areas. At A'q, the path 



