338 Upon the varl*€gated Colours of Bodies, ^c. 



Jigure so ingeniously traced by Newton in order to point out 

 the colours of the rings *. 



In the third place, the colours of the thinnest possible 

 pellicles have a very great liveliness ; those, on the contrary, 

 of the most intense coloured solutions are imperceptible 

 under an equally small thickness. It is for this reason that 

 the colouring of small plates of mica of an excessive tenuity 

 has no relation with the yellow colour of the mica in the 

 mass from which these leaves have been detached. They 

 are at that time in every respect similar to fragments of the 

 most colourless glass blown into bubbles, nearly of an equal 

 tenuity, and when mixed together we should have no sign 

 by which to distinguish them. 



Thus glass, mica, or any other substance, which under 

 a great tenuity are decked in the most brilliant colours, pass, 

 by increasing in thickness, to an absolutely colourless state, 

 or to a colour totally independent of those they exhibit when 

 in a thin state. 



But, it will be said, in order to assimilate a coloured mass 

 to an assemblage of parcels of a determinate thickness, it is 

 also requisite that these parcels be kept at a convenient di- 

 stance from each other. 



In this case, I answer, you will have a certain colour re- 

 flected and another colour transmitted precisely complemen- 

 tary to the first. Now, this double colouring never takes 

 place with respect to substances entirely diaphanous. 



The examples of the infusion of nephritic wood, and of 

 the precipitates of gold, are not applicable to this case, since 

 here, as I have previously shown, the colours reflected are 

 owing to molecules impermeable to light and disseminated 

 in a transparent liquid, and because we may alter the nature 

 of these molecules, or even have others of them in such a 

 manner as to change the colour reflected, without there be- 

 ing any change on that account in the colour transmitted. 



There has not hitherto been anv case known, which per- 

 mits us to consider a coloured body perfectly transparent, or 

 even dull, as composed of parcels of a determinate thickness,- 



* Optics, book ii. part ii. plate 2. fig. 6. 



and 



