256 
DR. A. A. RAMBAUT OX UNDERGROUND TEMPERATURE AT OXFORD 
therefore, be a transfer of heat which would render equation (e) no longer strictly 
applicable. 
A solution on the hypothesis that the heat which is conducted in this way may be 
represented by /rd leads to the two values 
x/v/k = 0-1189 from the annual wave, 
and v / ' 7r / /<: = 0T184 from the half-yearly wave, 
a very satisfactory agreement ; but the values found for ji k on the same hypothesis, 
viz. :— 
[jl/k = 0'0007 from the annual wave, 
and /j. k = 0'0080 from the half-yearly wave, 
differ too much to admit of any confidence in this additional term as representing 
the exchange of heat. 
We cannot, however, be much in error in taking 
0-1188 
as the value of v /7t//< for the gravel in which the thermometers are sunk. 
The value of the same quantity as found by Professor Everett from the Greenwich 
Observations was 
0-09175, 
and for the three stations at Edinburgh, from Professor Forbes’ observations, Lord 
Kelvin obtained, 
Calton Hill, trap rock.0*1154, 
Experimental Garden, sand . . . . "1098, 
Craigleith Quarry, sandstone . . . "06744. 
From the equations (/) we find from each thermometer a value of the amplitude 
and of the retardation of phase of each wave at the surface. 
Denoting the amplitude of the annual wave at the surface by P 01 , we have 
log P 01 = log Pj + . .r. * 
Substituting the values of Pj and x for each of the thermometers 2, 3. 4, and 5, we 
get four separate values of P 01 . These are 
Thermometer. P 01 . — ■ 
O 
5 . . . 15-48 
4 . , , -70 
3 . . . -70 
Mean = 15*65 = amplitude of annual wave at surface. 
* Cf. Quetelet, ‘ Annales de l’Observatoire Royal de Bruxelles,’ tome iv., 1845, p. 110. 
