690 



THE POPULAR SCIENCE MONTHLY. 



luminous index to travel several inclies, sliowing repulsion. A piece 

 of ice brought near causes the spot of light to travel as much in 

 the opposite direction. In order to insure the luminous index com- 

 'ui<f accurately back to zero, extreme precautions must be taken to 

 keep all extraneous radiation from acting on the torsion-balance. 

 The whole apparatus is closely packed round with a layer of cotton- 

 wool about six inches thick, and outside this is arranged a double 

 row of Winchester quart-bottles, filled with water, spaces only being 

 left for the radiation to fall on the balance and for the index ray 

 of lio-lit to get to and from the mirror. 



Fig. 5. 



However much the results may vary when the vacuum is imper- 

 fect, with an apparatus of this kind they always agree among them- 

 selves when the residual gas is reduced to the minimum possible; 

 and it is of no consequence what this residual gas is. Thus, start- 

 ing with the apparatus full of various vapors and gases, sucli as air, 

 carbonic acid, water, iodine, liydrogen, ammonia, etc., there is not 

 found, at the highest rarefaction, any difference in the results which 

 can be traced to the residual gas. A hydrogen vacuum appears the 

 same as a water or an iodine vacuum. 



The neutral point for a thin surface of pith being low, and that for 

 a moderately thick piece of platinum being high, it follows that at a 

 rarefaction intermediate between these two points pith will be re- 

 pelled, and that platinum will be attracted by the same beam of radi- 

 ation. This has been proved experimentally. An apparatus showing 

 simultaneous attraction and repulsion by the same ray of light is illus- 

 trated in Fig. 6. 



The pieces f g on the end of one beam consist' of platinum-foil 

 exposing a square centimetre of surface, while the extremities f g' 

 on the other beam consist of pith-plates of the same size. A wide 

 beam of radiation thrown in the centre of the tube on to the plates 

 g f causes g to be attracted and /' to be repelled, as shown hj the 



