STRAIN ON THE PHYSICAL PROPERTIES OF MATTER. 
819 
at various hours of the day. Sometimes the proportionate diminution of amplitude 
would seem quite as constant for different ranges of amplitude as could be expected, 
but at others there would be variations far exceeding anything which could be 
attributed to errors of observation. For some weeks I was quite at a loss to account 
for these variations, but at length was successful. Luckily, I had determined to 
make frequent observations of all three thermometers, and, whilst examining the 
contents of my note-book, it occurred to me that the greatest variations took place 
on those days and hours at which the variations of temperature were greatest. I also 
noticed that the logarithmic decrement became larger, whether the change of tempera¬ 
ture was of the nature of a rise or a fall. At first this explanation seemed almost 
incredible, for the variations of temperature would sometimes be only 1° C., or even 
less, and yet at the same time the value of the logarithmic decrement would be 
increased 30, 40, or even 100 per cent. My suspicions having been thus roused, 
I found it comparatively easy to verify them, and among several other experiments, 
made with the same object, tried the following :— 
Experiment VII. 
The same piano-steel wire as in Experiment III. The sun was shining rather faintly, 
and at intervals on that side of the house on which the wire was suspended, but before 
the end of the experiment passed over to the other side. 
Civil time. 
Temperature of the 
wire in 
degrees Centigrade. 
Alteration of 
temperature in 
degrees per hour. 
Time of 
100 vibrations 
in seconds. 
Logarithmic 
decrement due to 
internal friction. 
h. m. s. 
10 10 42 A.M. 
1033 
m. s. 
10 40 57 
9-33 
1-98- 
30+13-05 
•000636 
11 11 10 
10-00 
1*32-1- 
+14-35 
•000635 
11 41 24 
I TOO 
1-98 + 
+ 12-80 
•000753 
12 18 39 p.m. 
10-23 
0 . 
• • 
vibrator restarted 
12 48 52 
10-43 
0-396 + 
+ 12-90 
•000428 
1 49 15 
10-23 
0-198- 
+ 11-08 
•000356 
2 49 37 
10 10 
0-129- 
+ 10-93 
•000501 
4 50 21 
9 17 
0-461- 
+ 10-61 
•000702 
From 7 p.m. to 9 p.m., when the experiment terminated, the logarithmic decrement 
and the temperature became nearly constant, the value of the former being '000356. 
The changes of temperature, as shown in the second column, are very small, but we 
see what a large effect they have on the internal friction, which in one case became 
twice as great as when under a fairly steady temperature. The effects of fluctuations 
of temperature on the internal friction resemble those of mechanical jarrings, such as 
would be produced when the mass of the vibrator is altered, or even when the 
5 M 2 
