Vol.. 8, 1922 
PHYSICS: C. BARUS 
329 
1922 together, successive observations of the position of equiHbrium 
(about one hour apart) being inserted equidistant horizontally. Data 
for 1922 are distinguished by little circles with the nearest hour number 
of the observation attached. Data for 1921 are given in points, in corre- 
sponding positions, for easy comparison. F denotes that the attracting 
weight M on the right end of the needle is in front of it. R, that M is to 
the rear of the needle. Complete sets of observations (i.e. equilibrium 
curves F and R) are given between about 10 a.m. and 10 p.m. for each 
. year from July 24 to August 8 and to Aug. 15 in 1922. The exhaustion 
of the case in the morning and at night is shown on each day for 1922 only, 
as the data were not taken in 1921. In the same place C refers to cloudy, 
C to partly cloudy weather, 5 to sunshine or a clear day and R to rain. 
It is seen at a glance that the variations of the position of equilibrium y 
in the lapse of time, are of a different order in 1922, from their approxi- 
mate constancy in the given scale, in 1921. The contrast is startling with 
nothing easily apparent to account for it, unless it is traceable to a differ- 
ence in the respective va ua. For the exhaustions, as a whole, were thought 
to have been of about the same order of value in both years. In 1922 
there was fresh exhaustion on July 24, 28 and 30. Thereafter I left the 
apparatus with its slight leak ( 12 mm. per day), to itself, until August 
15. In 1921 there were no intermediate exhaustions and unfortunately 
the McLeod gauge was not attached to the case. 
The figure shows that all observations have a period of one day (24hr). 
Consequently the variations cannot be contributed by anything within 
the laboratory, though their enhancement is possible. They must in 
other words be originally meteorological, and due to solar radiation. Im- 
mediately after exhaustion (as on July 28, 30) there are apt to be large 
deflections, probably showing that the air within the case, though very 
rare, has nevertheless an appreciable cooling or other effect. For this 
reason I ceased to make exhaustions after July 30, and the observations, 
possibly for this contributory reason among others, gradually grew smoother 
and more normal. Any slight cooling of the inside of the case would tend 
to exaggerate the day time meteorological effect. If the cooling due to 
an exhaustion from 1 mm. to .001 mm. of pressure were instantaneous 
one would estimate that about .01 ca. would be abstracted from the air 
within the case. This is largely supplied of course by the massive case 
and the exhaustion is necessarily very gradual. Nevertheless the effect 
on the needle is always very marked. One cannot observe with the 
pump running, as the needle persistently clings to the wall of the case. 
An inspection of the equilibrium curves in July, 1922, shows that 
F and R are quite liable to intersect somewhere between 6 and 8 p.m. which 
jneans that the radiant forces are in excess (A^* negative) and they may 
