J.30 Astronomical and Nautical Collectidri», 



for in order that this effect may be produced, not only the- 

 two forces must be equal and in contrary directions, but they 

 must also act in the same right line, or be directly opposed, 

 to each other. 



The phenomenon of coloured rings, and that of the colours 

 developed by polarised light in crystallised plates, present a 

 particular case of interference, in which the undulations are 

 exactly parallel. But in the phenomena of diffraction, or in 

 the experiment with the two mirrors, which has been already 

 described, the rays which interfere always form sensible 

 though very small angles with each other. In these cases 

 the impulses to be combined with each other at the same 

 points, as belonging to the two systems of undulations, will also 

 act in directions forming sensible angles with each other : 

 but on account of the smallness of these angles, the result of 

 the two impulses is almost exactly equal to their sum, when 

 the impulses act in the same direction, and to their differ- 

 ence, when they are in contrary directions. Thus, in the 

 points of agreement or disagreement, the intensity of the 

 light will be the same as if the directions agreed more per- 

 fectly ; at least the nicest eye will not be able to discover 

 any difference in them. But although, with respect to the 

 intensity of the light, this case of interference resembles that 

 which has already been considered, there are other differ- 

 ences which modify the phenomenon very greatly, both with 

 respect to its general form, and to the circumstances neces- 

 sary for producing it. 



We may take, as a convenient example, the case of di^ 

 verging rays originating from the same luminous point, 

 and reflected by two mirrors slightly inclined to each other, 

 so as to produce two pencils meeting each other in a sen- 

 sible angle : the two systems of waves will then meet each 

 other with a slight inclination ; and it follows from this 

 obliquity, that if a semiundulation of the first system coin- 

 Ciides perfectly in one point with, a semiundulation of the 

 second, urging the fluid in the same direction, it must sepa- 

 rate from it to the right and left of the point of intersection* 

 and must coincide, a little further off, on one side with the 

 preceding semiundulation which is in a contrary direction^ 



