PROCEEDINGS 
OF THE 
NATIONAL ACADEMY OF SCIENCES 
Voume 6 JANUARY 15, 1920 Number 1 
AN EXAMPLE OF TORSIONAL VISCOUS RETROGRESSION 
By CarIv Barus 
Department of Physics, Brown University 
Communicated November 22, 1919 
In order to determine to v^hat degree it would be possible to eliminate 
radiant forces, I have recently installed a gravitational needle of the usual 
form, in a glass case, but vv^ith the balls at the end as heavy as the torsion 
wire would safely carry. The latter was hand-drawn (music) wire 0.022 
cm. in diameter and 154 cm. long, hanging from a torsion head attached 
to the pier. The balls were each over 1500 grams in mass and 29.3 cm. 
apart, between centers. The torsion coefficient of the system was found 
to be 154. The deflections of an externally attracting mass (also 1500 
grams) were read off with mirror and telescope at a distance of 290 cm. 
This makes the force to be measured F = 0.0095, if 5 is the deflection ob- 
served in cm. The period of the system would be about 14 minutes. 
As the two equal masses may be approached within less than 6 cm. of each 
other (between centers), a gravitational force of about 4.2 X 10 ~^ dynes 
is to be registered and this would be equivalent to a deflection of 5 = 0.47 
cm. Hence, by commutation, a total deflection of about 1 cm. may be 
expected, which could be much increased by increasing the masses {M = 
1500 g.) on the outside and inside of the case. On the quadratic inter- 
ferometer with achromatic fringes, the deflection 5 instanced, would be 
equivalent to a normal displacement of mirror AN = 0.023 cm., readable 
to 10~^ cm. directly, or to 4 X 10 ~^ per fringe. So far as mere observa- 
tion is concerned, the reading could be adequately made within 0.1%. 
I was in hopes, therefore, that the incipient motion of the needle, since 
air resistance, etc., is here negligible, could be studied as a case of uni- 
formly varied motion. 
Observations. — As these were made in a room, at first heated mildly 
and eventually more vigorously (see curve) they are useless for gravita- 
tional work. They are interesting, however, in their bearing on Maxwell's 
theory of viscosity, as they give evidence of a concealed torsional strain. 
