172 Trans. Acad. Sci. of St. Louis. 



fine wire attached to an insulated rod, which could be 

 laid upon the floor of the room, or lifted into contact with 

 a rod, one end of which rested upon one of the large 

 masses. The pinpoint terminal was meantime made to 

 face another grounded conductor. The suspended masses 

 then swayed to the other extreme position during the 

 next four minutes. By a repetition of this operation, 

 the arc of vibration could be increased from one or two 

 scale divisions to twenty-six in an interval of an hour. 

 The average of the extreme readings of consecutive vi- 

 brations, usually showed slow progressive changes of 

 two to five scale divisions per hour. Sometimes the 

 average reading showed a decrease and sometimes an in- 

 crease. The results were of the same order of magni- 

 tude whether the masses were connected with the nega- 

 tive or with the positive terminal of the machine, al- 

 though most of this work has been done by the positive 

 ' ' discharge. ' ' 



It was suspected that this result might be due to a 

 slight rise in the temperature of the walls of the metal 

 shield. If this were the case it must be assumed that 

 while the gap at the pin-point terminals was least, the 

 rise in temperature was least in that part of the sides 

 of the metal screen facing the large external masses. 



In order to examine this influence, a delicate air ther- 

 mometer was extemporized. The bulb consisted of an 

 oil-can having a capacity of 4.5 litres (one gallon). Con- 

 nected by a rubber hose with the nozzle of the can was a 

 horizontal tube of glass, having a length of 75 cm. and 

 an internal diameter of 0.76 mm. Within this tube was 

 a short column of water, serving as an index. The en- 

 tire device was insulated upon glass supports. A candle 

 flame 20 inches distant from the can caused the index to 

 move over a distance of 2 cm. in 15 minutes. When the 

 candle flame was removed and the can was electrified and 

 in electrical contact with the large masses and shield, no 

 change in the position of the index column could be de- 

 tected, when observed by means of a telescope. 



