under the Influence of the Electric Discharge. 377 



length ; but the results were not very constant. A few of 

 these observations are given in the following table. 



- 



+ 



V 



F=0. 



2. 



5. 10. 



14. 



P 



F=0. 



2. 



5. 



10. 



14. 



14-4 



0-74 



... |0-66| ... 





14-4 



0-50 





0-41 







6-2 



0-39 



0-33 ,0-43 0-42 



... 



6-2 



0-48 







0-26 





2-4 



0-34 



0-58 0-68 |0-46 





2-4 



017 



015 



0-25 



0-37 





1-2 



0-43 





0-59 ,0-77 090 



1-2 



0-08 





0-30 





050 



X 



1-26 



lib 



1-4 1-35 | ... 



X 



0-06 



0-32 



0-46 



0-91 





The constantly observed fact is to be particularly mentioned, 

 that at low pressures the heating at the positive electrode is un- 

 usually small when no spark is employed, but that it increases 

 very rapidly as soon as a spark is introduced. 



We may bring together the results of the observations of 

 the heat produced when no spark is included in the circuit as 

 follows : — 



(1) The total heating first decreases, and then rapidly in- 

 creases, as the pressure decreases. 



(2) The heating of the tube decreases rapidly, and then in- 

 creases very slightly. 



(3) The heating at the positive electrode first decreases ra- 

 pidly, and then increases slightly. 



(4) The heating at the negative electrode first decreases slowly, 

 and then increases rapidly. 



Hence it follows that, for the thermal phenomena in the 

 whole discharge-tube, in the experiments which I have made, 

 and under normal conditions without the use of any air-spark, 

 the phenomena at the negative electrode are the most im- 

 portant. 



But from experiments on the positive discharge it follows, 

 further, that deviations from the laws stated above depend 

 upon a peculiar behaviour (p. 374) of positive electricity, which 

 becomes specially important, not when the passage of electricity 

 through the tube takes place at the normal potential corre- 

 sponding to the construction of the tube, but when by the use 

 of sparks or external resistances the quantity of electricity 

 passing in each discharge is increased. 



4. Absolute Determinations. 



In order to obtain a measure of the quantity of heat pro- 

 duced in the capillary tube per unit length by the unit quan- 

 tity of electricity, special measurements were made. 



The calorimeter was of glass, the fluid in it water. The 

 length of the portion of the capillary tube surrounded by 



