Repulsion resulting from Radiation. 147 



desired colours can be measured either against the total radiation 

 from a candle, its luminous rays, or any desired colour. One 

 coloured ray can be balanced against another coloured ray by having 

 differently coloured screens on either side. 



The variations in the luminosity of a " standard " candle will 

 cease to be of importance. Any candle may be taken ; and if it be 

 placed at such a distance from the apparatus that it will give a 

 uniform deflection, say of 100 divisions, the standard can be repro- 

 duced at any subsequent time ; and the burning of the candle may 

 be tested during the photometric experiments by taking the deflection 

 it causes from time to time, and altering its distance, if needed, to 

 keep the deflection at 100 divisions. 



If the pith bar in this instrument be blacked on alternate halves, 

 an impetus given by a ray of light always acts in the same 

 direction of movement. A candle causes it to spin round very 

 rapidly until the suspending fibre is twisted up, and the rotation is 

 stopped by the accumulated torsion. 



By arranging the apparatus so that the black and white surfaces 

 are suspended on a pivot instead of by a silk fibre, the interfering 

 action of torsion is removed, and the instrument will rotate con- 

 tinuously under the influence of radiation. To this instrument 

 the author has given the name of the " E-adiometer," or " Light- 

 Mill." It consists of four arms of very fine glass, supported in 

 the centre by a needle-point, and having at the extremities thin 

 disks of pith lampblacked on one side, the black surfaces all facing 

 the same way. The needle stands in a glass cup ; and the arms 

 and disks are delicately balanced, so as to revoh T e with the slightest 

 impetus. 



In the 'Proceedings of the Eoyal Society' for 1875 (vol. xxiii. 

 p. 373), the author gave a brief account of some of the earlier 

 experiments with these instruments. In the present paper he 

 enters very fully into the various phenomena presented by them, 

 and gives Tables showing the number of revolutions made by the 

 radiometer when exposed to a constant source of light removed 

 different distances from the instrument. The law is that the 

 rapidity of revolution is inversely as the square of the distance 

 between the light and the instrument. 



When exposed to different numbers of candles at the same 

 distance off, the number of revolutions in a given time is in 

 proportion to the number of candles, two candles giving twice the 

 rapidity of one candle, and three giving three times, &c. 



The position of the light in the horizontal plane of the instru- 

 ment is of no consequence, provided the distance is not altered ; 

 thus two candles one foot off give the same number of revolutions 

 per second, whether they are side by side or opposite to each other. 

 From this it follows that if the radiometer is brought into a uni- 

 formly lighted space it will continue to revolve, This is proved 

 to be the case by experiment. 



The speed with which a sensitive radiometer will revolve in full 

 sunshine is almost incredible ; nothing is apparent but an unde- 

 fined nebulous ring, which becomes at times almost invisible 



L2 



