266 



Mr. J. Barton on the Injiexion of Light. 



Indeed, these values are so much at variance with one another 

 as to destroy all confidence in the theory ; since the differences 

 are far greater than can be accounted for by any supposable 

 errors of observation, even had the observer been less remark- 

 able for accuracy than Newton. This will appear clearly from 

 the following Table, in which I have taken the length of an un- 

 dulation such as it is given by the first of the above observa- 

 tions, and then computed by Fresnel's method the distances of 

 the knives from the paper in each of the succeeding observa- 

 tions. 



Distances between the Distances of the knives from the paper 

 edges of the 

 knives. 

 •012 

 •020 

 •034 

 •057 

 •081 

 •087 



Still more are these observations at variance with Young's 

 theory, which supposes the phenomena to depend on the in- 

 terference of the two rays reflected from the edges of the slit. 

 In this case the centre of the spectrum would be always bright, 

 the lengths of those two rays being equal. 



M. Biot's observations were made in red light*. He has 

 not mentioned the distance of the slit from the opening through 

 which the light entered; but as the length of an undulation in 

 red light is given by M. Fresnel, we may reverse the calcu- 

 lation, and compute what this distance must have been from 

 the other data. It will be seen that the results given by the 

 theory are impossible. 



In the first three observations, the distance of the slit from 

 the opening in the window-shutter, as computed by Fresnel's 

 hypothesis, here turns out to be a negative quantity. In other 

 words, this distance is greater than infinity I 



It may be proper briefly to state the method by which these 



Trait c de 2 3 /ti/siqiie> torn. iv. p. 757. 



f Ibid, page 764. 



