Nipher—Gravitational Repulsion. 189 
contact was made. A maximum reading was obtained at 
11:18 a. m. and the masses swayed in the opposite direc- 
tion during the next 14 minutes. Forced vibrations were 
again impressed upon the suspended masses, but they 
were less violent in character than those at the close of 
the observations represented in Plate XLVI. At 12 ob- 
servations were discontinued until 12:48 p. m., when the 
reading showed only a slight decrease. At 1 :00 p. m. the 
terminals were reversed and the direct contact was re- 
moved. The conditions then were as represented in Fig. 
3. At this time the apparent decrease in the initial at- 
traction which existed at the beginning of the observa- 
tions on June 2, was about 340 per cent. In other words 
the apparent repulsion was then more than twice the 
initial attraction. This conclusion seems to be fully jus- 
tified by the amazing increase in the attraction which at 
once resulted. At 1:53 p. m. direct contact between the 
pin-point conductors and the shield and large masses was 
made by dropping a wire into position. It was removed 
at 2:02 p. m. Forced vibrations were again impressed 
upon the suspended masses. The machine was stopped 
at 2:19 p. m. and I was called away for an hour. A few 
readings taken between 3:12 and 3:22 p. m. shown at the 
close of Plate XLVII indicate that the attraction had 
then approached closely to the initial value on the morn- 
ing of June 2. 
It will of course be understood that no attempt has 
been made to secure results of precision in this work. 
The only aim has been to determine whether or not it 
would be justifiable to construct the much more expen- 
Sive apparatus which will be required for such results. 
The suspended masses must be capable of being electri- 
fied independently and the enclosing walls must surround 
them in symmetrical form, so that their inductive action 
will not produce deflection of the suspended masses. If 
they are suspended in highly rarefied air, it may be neces- 
sary to use a metal wire rather than a quartz fiber, which 
must then be attached to an insulated torsion head. A 
