70 



KENNELLY-SANBORN— INFLUENCE OF 



[April 24, 



from the test-wire when held stationary in the air. Professor Mor- 

 ris's method of graphic representation has the advantage that it indi- 

 cates directly the magnitude of the empirical correction v^. If we 

 take ^^0^30, we have from (4) 



■ = kVv -f- 30 



abwatts per cm. 



(5) 



Figs. 9 and 10 give the graphs of Pc/Q on logarithm-paper, for 

 the various sets of observations. It will be seen the observations lie 

 not far from straight lines. These lines have a gradient of i :2, or 

 correspond to a square-root law, or exponent of 1/2 as in (5) ; ex- 

 cept for the vacua (0.44 and 0.66 megabar), where the gradient is 

 approximately 4:10; or would more nearly indicate a relation 



P 



j = k{v-i- sor 



abwatts per cm 



'C. 



- (6) 



Fig. 10. Graphs of (-^J against 7^ + 30 on logarithm paper for <9 = 538° C. 



