198 Dr. a. Hodgkinson oh 



though the cause varies, the effect, so far as regards tint and 

 sequence, under var}-ing incidence of Hght is the same. Hence 

 the reason why some observers have inferred thin plates as 

 the cause, others fine Hnes, in the same object. Though 

 identical in tint and sequence of colours, it occurred to me 

 to ascertain whether there were not some features in the 

 colour phenomena of fine lines which differed from the 

 colour phenomena of thin plates, because, if such could be 

 found, the two phenomena could be distinguished and the 

 correct structural cause inferred. To answer this question — 

 and I go into detail on this point to serve as an example of 

 methods I have adopted, with the necessary modifications, 

 in other instances — I took examples of both these struc- 

 tures, thin plates and fine lines, prepared artificially, so 

 that there could be no doubt of their structural nature. For 

 a thin plate I took an iridescent film of mica, for fine lines 

 a small diffraction plate consisting of a series of fine lines 

 engraved on silvered glass by Zeiss. These I examined in 

 the following manner : — First by transmitted light ; placing 

 the diffraction grating, the lines of which were of course suffi- 

 ciently near to one another to produce diffraction colours, on 

 the stage of the microscope, I illuminated it from below in the 

 usual manner, except that I used a diaphragm with an 

 aperture made by a fine needle point. Employing a low 

 power ( I in.), I focussed, not for the grating but for this 

 aperture, so arranging the grating that the light passed 

 through it before entering the objective. On observing the 

 result, as seen through the eyepiece, a central colonrless^ 

 image of the aperture of the diaphragm is seen, and on 

 either side of this central image, and in a plane at right 

 angles to the direction of the markings, is a series (in this 

 instance two) of spectra of this opening in the diaphragm 

 having the violet end of each spectrum towards the central 

 opening. If the grating be rotated in altitude on an axis in 

 the direction of the Hnes, the colourless imaee still retains 



