182 
IOWA ACADEMY OP SCIENCE 
case upon the zero readings of the balance. One set of results must suffice 
here. In these, the conditions were in a measure controlled by having an 
eleqtrie light mounted outside the case at the left end, and turning it on and 
off at will. The accompanying figure best illustrates the results, which are 
tabulated below. 
Time P. M. 
4:10 
4:17 
4:20 
4:23 
4:20 
4:34 
4:30 
4:37 
4:39 
4:42 
4:45 
4:48 
4:52 
4: 54 
4:58 
5:03 
5:05 
5:07 
5:10 
5:12 
5:17 
5:19 
5:21 
Galv. Read, j Pointer Read. 
Remarks 
2.9 
15.1 
16.9 
17.9 
11.0 
15.0 
16.8 
18.0 
19.0 
19.5 
20.1 
20.5 
20.9 
11.0 
5.3 
4.9 
4.3 
4.1 
10.12 
10.47 
10.55 
10.69 
10.60 
10.30 
9.97 
9.84 
10.04 
10.21 
10.30 
10.35 
10.32 
10.29 
10.20 
9.84 
9.33 
9.24 
9.01 
9.01 
Liglit off 
Light on 
Light on 
Light on 
Light off 
Light off 
Light off 
Light on 
Light on 
Light on 
Light on 
Light on 
Light on 
Light on 
Light on 
Light off 
Light off 
Light off 
Light off 
Light off 
Light off 
Light off 
Light off 
An examination of these curves suggests several things. Note the remark- 
able promptness of the pointer reading effect when the light is turned on. 
The second drop in the pointer reading precedes the drop in temperature 
difference, and is gradual instead of sudden, showing that it is due partly to 
other causes. This does not occur in the first and shorter exposure to radia- 
tion. No doubt the other causes operating to bring down the pointer reading 
are the equalization of temperature in the beam by conduction and the gradual 
growth of air currents in the case by contact with the warm glass on the left 
side. The same thing is noticed in the rise of the pointer reading at the end 
of the series, while the temperature difference is still decreasing. It is to be 
noted that the whole range of temperature difference is less than one degree. 
It is evident that air currents in the case, set up by difference of tempera- 
ture at the two ends, have an effect opposite to that of expansion in the beam. 
For. the current is downward at the cooler end, bearing down on the pan; 
while the corresponding arm is the shorter, which tends to raise the pan. It 
is difficult to say how much the effect of expansion in the beam may be thus 
modified by the air currents; but it is likely that with broad pans the modifica- 
tion may be considerable, providing the air currents have had sufficient time 
to get fairly started. 
The statement is made in some laboratory manuals that the difference in 
length between the two arms of a first-class balance when in adjustment will 
be found negligible for all but the most precise weighing, for which latter pur- 
pose the interchange of object and weights is recommended. This is perhaps 
quite true, providing the beam is at a uniform temperature; but the experi- 
ments herein described show how marked may be the effect of even slight 
temperature inequalities. Moreover, if such inequalities exist, unless they can 
