348 On the Longitudinal Lines of the Solar Spectrum. 



slit through which the light is admitted into the dark room, or 

 to other similar causes ; and that they were in no way related 

 to the constitution of the light itself. 



The numerous experiments which I have made in connexion 

 with this subject, have led me to the conviction that the longi- 

 tudinal lines are not due to the irregularities alluded to, but are 

 produced by interference. Whoever is accustomed to experi- 

 ments on light, will find the mere inspection of these lines suffi- 

 cient to convince him that they are due to no mechanical cause. 

 The clearness and beauty with which they exhibit themselves, 

 and their sharp and definite character throughout their entire 

 length, distinguish them at a glance from those which might be 

 produced by unevenness of the slit's edge, particles of dust, im- 

 perfections of the apparatus, and so forth. 



In the first place, I have observed that the longitudinal lines 

 are entirely absent when a large lens is not applied, and when it 

 is placed close to the prism and at right angles to the rays issuing 

 from the same. I have further seen, that the lens changes the 

 breadth of the spectrum only, and not its length. 



Thus in one of my experiments, which was conducted with an 

 equilateral vertical prism and a biconvex lens of 90 centimetres 

 focal distance, after ascertaining by trial the position in which 

 the spectrum was most clearlyshown, I found its dimensions to be — 



Length . . . 134 centims. 

 Breadth ... 3-2 ... 



The lens was then removed, and the position of the screen and 

 prism remaining unchanged, the dimensions were found to be — 



Length 13*4 centims. 



Breadth . . . 15-8 ... 



Hence the introduction of the lens caused the disappearance 

 of 12"6 out of 15 - 8 parts of the spectrum; the light must have 

 been compressed from a space of 15"8 to a space of 3 - 2. 



In the latter space, the rays which had passed through the 

 lens overlaid each other, as may be rendered evident by a very 

 simple experiment. It is only necessary to move a bit of card- 

 board close to the lens from top to bottom, or the reverse, and 

 thus to receive a portion of the rays passing through it. It is 

 then seen, that no matter how great the portion may be which 

 is thus intercepted, the dimensions of the spectrum remain un- 

 altered, its brightness alone being more and more diminished as 

 the intercepted portion becomes greater. This experiment esta- 

 blishes the fact of superposition, and the production of the lon- 

 gitudinal lines by interference is a simple consequence of this. 



Let ST, Plate IX. fig. 7, be the breadth of the spectrum on 

 the surface of the lens, and sey the plane of projection at the 



