•4-t Proceedings of the Royal Irish Academy. 



We cannot fully explain the want of steadiness in the numbers for 

 uegative charge as compared with those for positive charge^ but there is one 

 factor which should make a decided difference in the form of the curves 

 connecting production and speed according a* the metal surface is positive 

 or negative. Later work I below showed that an increase of tempe- 



rature of the metal cap decreased the rate of charging, provided the charge 

 tive, but increased the rate of production of a negative charge on the 

 metal. An inert mperature tends to make the metal surface acquire 



a negative charge. As an increased speed of rubbing means an increase of 

 temperature, this eHect would tend to flatten the curve and cause the pro- 

 duction to reach a maximum when the metal was positive, while it would tend 

 to cause a more and more rapid production at high speeds when the charge 

 was negative. 



In addition to this temperature effect we must remember that at high 

 speeds the rapid return of the metal over the same silk surface brings it into 

 contact with an oppositely electrified surface. This Bhould flatten the curves 

 connecting the i _ ug with the speed whether the charge on the 



metal is positive or negative. When the metal is charging positively, we 

 have, thi two effect* n th>' Bame direction and diminishing the 



production of charge at high speeds, but acting in opposite 

 ■ ions when the metal is ch irging 

 This reasoning does not, however, explain the variable nature of the 

 tits we obtained when the metal was _ _ gatively. These measure- 

 ments were among the first made, and before we had learned Borne of the 

 conditions m ■ Its. 



I • tain a posith harge on the metal when rubbed with silk it is 



well to have the metal surface polished. When polished, some metals 



_• and behave very steadily. A clean surface of the same 



metal free from polish may ch _ itively, and clearly iucon.-i.-U-nt results 



may be obtained at a transition .-lage. 



To return to the results with the metal Burface charging positively, the 



curves would require correction for both the effects mentioned above in order 



irrive at the true relation between rate ol production and Bpeed, but there 



i> no doubt that the curves tend to a maximum a.- tin- speed is incrcas 



I mi similar curves for neg _ 11 tin- metal it will be m-eessary 



t" pre] Standard way without actually polishing it. 



We b 1 direct experiments with metal surface.- polished and 



then rubbed very .-lightly with emery paper. It in most cases was 



that contact with the emery papej I to make the metal charge nega- 



tively. If charging positively when polished, it gave either a smaller positive 



