246 
DR A. A. RAMBAUT ON UNDERGROUND TEMPERATURE AT OXFORD 
We thus obtain the following table for the correction from the platinum to the air 
scale, for every degree of the former from —15° to +25°. 
Table Y.—Reduction from the Platinum to the Air Scale. 
■pt. 
Corr. to air. 
pt. 
Corr. to air. 
pt. 
Corr. to air. 
pt. 
Corr. to air. 
o 
-15 
o 
+ 0-256 
o 
-5 
o 
+ 0-078 
o 
+ 5 
O 
-0-071 
o 
+ 15 
o 
- 0T91 
14 
•237 
4 
■062 
6 
•084 
16 
•201 
13 
•218 
3 
•046 
7 
•097 
17 
•211 
12 
T99 
2 
•030 
8 
•no 
18 
•221 
11 
T81 
-1 
+ -015 
9 
T22 
19 
•231 
10 
T63 
0 
•000 
10 
•134 
20 
•240 
9 
T46 
+ 1 
- -015 
11 
T46 
21 
•249 
8 
. T28 
2 
•029 
12 
T58 
22 
•257 
7 
•111 
3 
•043 
13 
•169 
23 
•266 
6 
•095 
4 
•057 
14 
T80 
24 
•274 
- 5 
+ -078 
+ 5 
- -071 
+ 15 
- T91 
+ 25 
- -281 
The reduction of the observations is thus of a very simple character ; but it may be 
still further simplified, and the chance of arithmetical errors occurring in individual 
cases greatly reduced, if not wholly removed, by the preparation of a table giving 
the total correction to the bridge wire reading for each arrangement of coils, the 
arguments in each case being the bridge wire reading and the temperature of the 
box. Of course, if a great variety of coils were in use, such tables would attain 
dimensions out of proportion to their usefulness, as it would be necessary to construct 
a special table for each separate combination of coils. But in observations of 
underground temperature, the range of the readings of an} 7 particular thermometer 
is comparatively limited, so that only two different arrangements of the plugs are 
necessary, coil A serving for about eight months of the year, and coil B coming into 
use for about four months in summer. 
It is thus a simple matter to compute tables that will cover all cases. Tables of 
this sort were prepared for each of the earth thermometers, from which the correction 
to the bridge wire reading, to reduce to the corresponding temperature on the air scale, 
could be obtained at one step by simple interpolation. 
For some little time before and after the epochs at which the change from one coil 
to the other is made (about the end of May and the end of September) it is possible 
to read the thermometers with either arrangement, and thus a check can be imposed 
upon the general performance of the apparatus. 
In 1899 the changes were made on June 1 and September 26, and the following 
readings of the 10-foot thermometer, in which the changes are very slow and regular,, 
afford very satisfactory evidence of the consistency of the readings. 
