SUPERPOSITION OF PARAGENIC SPECTRA, ETC. 223 



binations of 1000 with 1000, 1000 with 2000, 1000 with 500, 2000 with 500, and 

 in the combination of four surfaces of 2000, 1000, 100, and 500. 



In the combination of 1000 and 500, and in no other, a very peculiar system 

 of bands is seen with a lens. They are not serrated, and not perpendicular to 

 the grooves. The system consists of two sets equally inclined to the direction of 

 the grooves, when the grooves in one plate are slightly inclined to those in 

 the other. By diminishing the inclination of the grooves, the inclination of the 

 bands to the direction of the grooves diminishes, and when the grooves become 

 parallel, the bands become parallel and disappear. 



These bands must have a different origin from those previously described, as 

 they are similar in number upon all the prismatic images. 



In these experiments the duplication of the bands on the second spectrum, and 

 their increase in arithmetical progression on the other spectra, is a remarkable 

 fact which it is difficult to explain. The second spectrum differs only from the 

 first, and the third from the second, only in their length ; and we can hardly 

 suppose that they have a property in a direction perpendicular to their length, 

 or to Fraunhofer's lines, which would increase the number of their bands. 



The bands which we have described are more distinct when the spectra are 

 pure or formed from a narrow line or bar of light ; but when we wish to see the 

 bands on the bar of light or the central image O (fig. 4), the spectra must be 

 formed from wide spaces which gave impure spectra. 



In order to examine the interference bands under different conditions, I placed 

 (as in fig. 6) a plate of polished steel at different distances from another plate 

 of steel, containing six systems of grooves executed by Sir John Barton, varying 

 from 312-5 divisions in an inch to 10,000. When the light was reflected twice 

 from the grooved surface and once from the plain steel surface, the bands which 

 covered the colourless image and the paragenic spectra were splendid beyond 

 description, and unlike anything of the kind that I had previously seen. 



1. The bands were parallel to the grooves, or to the lines in the spectra. 



2. They are smaller and more numerous when the grooves are wider or fewer 

 in an inch. 



3. They become smaller and more numerous when the distance of the plates 

 is increased. 



4. They are smaller and more numerous when the angle of incidence is 

 increased. 



5. They become more numerous by increasing the number of reflexions. 



6. They appear like minute black lines upon the colourless image, but when 

 their magnitude is increased, they appear like blue or pink bands on a ground of 

 a different colour, which is generally white or whitish blue. 



These bands were visible on the systems of grooves, 312-5, 625, 1250, and 

 2500 in an inch, but not on the systems of 5000 or 10,000 in an inch. 



