180 MKUSCHEL ON COLOURED RINGS. 



a set that is seen by transmission, and it will therefore luive a 

 white centre. The rays 1, 2*, 4, passing through the point of 

 contact, will also form a transmitted set with a white centre, 

 which will be seen when the reflection from 4 to 5 conveys it 

 to the eye. Hut these two sets have no connection with each 

 other; and as primary sets are independent of all other sets, 

 I have only to prove, that this secondary set belongs not to 

 the primary one which is seen, but to another invisible one. 

 This may be done as follows. 



Introduce the black strip of card, that has been mentioned 

 before, till it covers the mirror at 7; this will takeaway the 

 strong reflection of light, which overpowers the feeble illu- 

 mination of the rays 1, 2, 3 ; and the real hitherto eclipsed 

 primary set, belonging to the secondary one with a white cen- 

 tre, will instantly make its appearance with a black one. We 

 may alternately withdraw and introduce again the strip of 

 card, and the centre of the primary set will be as often 

 changed from one colour to its opposite ; but the secondary 

 set, not being dependent on the rays 6, 7, will not be in the 

 least affected by the change. 



If the contact should have been such as to give both sets 

 with orange centres, the introduction of the strip of card will 

 prove, that the set which is primary to the other has really a 

 green centre. 



Another way of destroying the illusion is to expose the 

 same arrangement to a brighter light, and at the same time 

 to increase the obliquity of the angle of incidence; this 

 will give a sufficient reflection from the surface of the glas* 

 to be no longer subject to the former deceptive appearance; 

 for now the centre of the primary set will be black, as it 

 ought to be. 



XXII. Of the reflecting Surfaces. 



Situation of The ra > s °f Kjfht, tnat *° rm rings between glasses, must 



the surfaces undergo certain modifications by some of the surfaces 

 £l ,eflcct the through which they pass, or from which they are reflected; 

 aud to tind out the nature of these modifications, it will be 

 necessary to examine which surfaces are efficient. As we 

 see rings by reflection, and also by transmission, 1 shall begin 

 with the most simple, and ^how experimentally the situation 



of 



