GRAVITATION AS AFFECTED BY TEMPERATURE. 
379 
heating coils are substituted. When, after say two hours, the lead spheres are at a 
high temperature, the thermometers are again substituted for the coils, so that for 
the set of hot readings all the conditions are exactly the same as for the cold readings, 
201-3 
201-2 
201-1 
If 
£ 
c 
w 201-0 
UJ 
O 
Z 
< 
cz 
200-9 
200-8 
Fig. 13. Graph for Table II. The crosses show the first series of readings, the numbers being in order. 
The circles show the second series. 
except as regards the raised temperature in masses M, M. A graph is given for the 
result. The effect is 
+ 1-2/1 0 5 per 1° C. 
Table TIL is another instance. It differs in some details from the preceding table. 
In particular, the fourth column consists of only two entries at each A, B position. 
These are two consecutive extreme scale readings. Call them a, b. Then if Z be the 
rest position, we have, supposing geometrical progression in the amplitudes, 
2 _ ad b 
1 + d 
where d — decrement (i.e., ratio of one swing to the preceding one). 
The value of d is found independently before or after the experiment. The method 
used here has several advantages over the usual method of taking three or more 
extreme scale readings, viz. :— 
(1) The experiment is shortened. It may be very tedious when, as here, each half¬ 
swing takes 160 secs., and the whole series lasts 6 to 8 hours. 
(2) It is important when the large spheres are hot that they should remain as 
short a time as possible at any one part of the vacuum tube. We reduce the time of 
rest at any place by this process to about two-thirds of the least time possible by any 
other means. 
2 
20 ° 40 ° 60 ° 80 ° 100 ° 120 ° | 40 ° 160 ° 180 ° 
T EMP. ( Cent i grade). 
