THE MECHANICAL ACTION OF LIGHT. 693 



so on up to twenty-four candles. A small radiometer was found to 

 revolve at the velocities shown in the following table, when exposed 

 to the radiation of a standard candle five inches off: 



Time required for One JRevolution. 



Source of Radiation. Time in Seconds. 



1 candle, 5 inches off, behind green glass 40 



" 5 " " blue " 38 



" 5 " " purple " 28 



" 5 " " orange " 26 



" 5 " " yellow " 21 



" 5 " " light-red".... 20 



In diffused daylight the velocity was one revolution in from 1.7 sec- 

 onds to 2.3 seconds, according to the intensity of the incident rays. In 

 full sunshine, at 10 a. m., it revolved once in 0.3 second, and at 2 p. m. 

 once in 0.25 second. 



When heat is cut off by allowing the radiation to pass through a 

 thick plate of alum, the velocity of rotation is somewhat slower, and 

 when only dark heat is allowed to fall on tlie arms (as from a vessel 

 of boiling water) no rotation whatever is produced. 



In all respects, therefore, it is seen that the radiometer gives in- 

 dications in strict accordance with theory. 



Several radiometers, of various constructions as regards details, 

 but all depending on the above-named discovery, have been exhibited 

 at the Royal Society, where their novelty and unexpected indications 

 excited a considerable amount of interest. 



This form of instrument is of too recent a construction for me to 

 be able to do more than draw brief attention to a few of the many 

 uses for which it is applicable. 



By timing the revolutions of the instrument when exposed direct 

 to a source of light a candle, for instance the total radiation is 

 measured. If a sci-een of alum is now interposed, the influence of heat 

 is almost entirely cut off, the velocity becomes proportionately less, 

 and the instrument becomes a photometer. By its means photometry 

 becomes much simplified; flames the most diverse may readily be 

 compared between themselves or with other sources of light ; a " stand- 

 ard candle" can now be defined as one which at x inches off causes 

 the radiometer to perform y revolutions per minute, the values of x 

 and y having previously been determined by comparison with some 

 ascertained standard ; and the statement that a gas-light is equal to 

 so many candles may, with more accuracy, be replaced by saying that 

 it produces so many revolutions. 



To photographers the radiometer will be invaluable. As it will 

 revolve behind the orange-colored glass used for admitting light into 

 the so-called dark-room, it is only necessary to place one of these in- 

 struments in the window to enable the operator to see whether the 



