252 On the Application of Photography to Diffraction. 



other figures used in Sir John Herschel's experiments within the 

 space of the pupil of the eye on a collodion plate, that series 

 of beautiful phfenoraena might be produced without the employ- 

 ment of a telescope, or at any rate by the employment of a tele- 

 scope of veiy low power. 



Several circumstances tend to prevent perfect accuracy, and 

 claim attention, either to be obviated as we best can, or to be 

 dismissed as incapable of doing material harm. 



1 . Inequality in the surface and in the substance of the glass 

 which is to receive the collodion film. 



2. Inequality in the collodion film ; in its thickness from 

 poiu-ing off, or wrinkles on its surface left by the evaporation of 

 the sether. 



3. The difficulty of placing the glass to receive the image in 

 a plane exactly pai'allel to that of the object. 



With respect to the second difficulty, it is to be considered 

 that the thickness of the collodion film is of an order quite com- 

 parable with the length of a wave of light ; in fact, I found that 

 the film of an average collodion from a hundred square inches, 

 when dried and lightly pressed, occupied considerably less than 

 ^^^th of a cubic inch, so that the thickness was certainly not 

 more (and probably very much less) than g^'y^dth of an inch. 

 The inequality arising from wrinkles is a small fraction of this. 

 The inequality arising from pouring off must also, for the small 

 surface I would use, be a very small fraction of the whole thick- 

 ness ; and such as it is, approximates to that of a very acute- 

 angled prism, whose effect would be quite inappreciable. By 

 cementing a piece of thin glass over the surface of the picture, 

 it is probable that the effect of the wrinkles would be neutralized, 

 at the same time that protection would be afforded to the film. 



The third source of error is the most important, but is moi-e 

 in the power of care and delicacy than the others. To find the 

 effect of a small error of this sort in displacing the points of the 

 figure, let ab be the correct, and AB 

 the actual position of the surface of 

 the film, C the centre of the picture, 

 of the lens. Then the point v/hich 

 shoiUd be at the distance Op from C 

 is at the distance CP. 6 ^ 



N«^^' CP-C;,= ggxAAC«. 



If we are to have, say 500 lines in Cjo = i of an inch, the en-or 

 ACfl, which would remove the 500th line to the place of the 

 501st, would be ^^, or rather more than 2° for a focal length of 

 10 inches. 



