76 



THE ROYAL SOCIETY OF CANADA 



that the instrument behaved as a radiometer, the Speed of Rotation 

 being proportional to the intensity of radiation. 



The agreement of the points with the curve was, however, 

 much better for large values of D and T than for small values, the 

 points for small values lying on a curve which does not go through 

 the origin. The small values of D and T represent large intensities 

 and speeds and to show them more clearly on a graph I have taken 

 Crookes' data and plotted Revolutions per minute (R.P.M. or N.) 

 as abscissae and the inverse square of the Distance as ordinates. 

 Fig. 1 gives the plot. The upper inset graph shows that for small 

 intensities and speeds the inverse law held and that the R.P.M. 



were proportional to the intensity. The larger graph departs from this 

 straight line (0 B in the large graph corresponds to ^ in the small 

 one. 



For large speeds (small distances) the graph is far from straight, 

 the speeds being less in proportion than the corresponding value 

 1 



of 



Z)2 



Either for great speeds the R.P.M. are not proportional to 



the radiation or for such small distances we cannot assume that the 

 radiation from a candle is proportional to — . 



For his Circle of Candles experiment, Crookes plots Times of 

 Rotation as abscissae and Number of Candles as ordinates and gets 

 practically a hyperbola. 



Replotting with R.P.M. as abscissae and number of candles 

 as ordinates, I get Fig. 2, which shows a good average straight line 



