34:4 C. Bar us — Effect of Temperature and of Moisture 



of the rise of temperature of nearly 100°. After repeated 

 trials, however, I found that the high temperature so far 

 diminished the insulation of the hard rubber bushings of the 

 condenser, that no measurements would be trustworthy. On 

 cooling the condenser, the insulation again became perfect. 

 Quartz insulators suggest themselves as probably alone avail- 

 able. 



23. Comparison of old and new grids. — Trial was made 

 with freshly cut phosphorus, the grids used in the above work 

 being as much as a year old. The results show a like order of 

 values for both, in spite of the intense fuming of the new 

 grid. Nevertheless the latter is apt to be from two to three times 

 stronger than the old grid. Thus after thoroughly drying the 

 new grid over calcic chloride and testing it at 21*5°, the data 

 were 



dV/dt =z -42, (ds/dt)Js a =-20, 



•43, =-20. 



Exposure to the air of the room for about half an hour after 

 desiccation was without effect. 



This experiment proves that the air in all the earlier experi- 

 ments was undersaturated, agreeing with an earlier paper (1. c) 

 from which values are quoted in the next paragraph. 



24. Older data compared with the present data. — It is finally 

 worth while to adduce the corresponding data. Table YI, of an 

 earlier paper. The equivalent colors of the steam tube are the 

 yellows and crimsons of the second order. The temperatures 

 are 20°-30°. 



Table VI. — Relevant data from earlier papers, E a — 40 volts. 



Color. Temperature,*?. WxdV/dt 10 2 x (ds/dt) /s 



Yellow, 30° 50 ' 16 



Yellow 2 30 65 17 



Crimson, 30 43 11 



Yellow-green, 21 50 13 



Yellow, 24 58 22 



Crimson,. _ 24 42 14 



Clearly then, for dV/dt = *43, nearly, the order of values 

 for the currents is the same here (the experiments made over a 

 year ago but with the same grids) as in the above experi- 

 ments. The room being very hot insured dryness in that 

 work without preliminary desiccation. Large volumes of dry 

 air were passed over the grids, moreover, for colors of the first 

 order were principally observed, requiring even as much as 

 3 liters /min. per observation. 



The arrows (w, new and 0, old) in figure 9 show the relation 

 of the results in the last table to the values for freshly cut 

 phosphorus in excess given at the beginning of paragraph 23. 



