420 PROFESSOR W. THOMSON ON THE REDUCTION OF 
units of length, but in terms of thicknesses corresponding to equal conducting — 
" powers and thermal capacities, and if we continue to designate the thickness of 
the first stratum by its number 3 of French feet, our reckoning for the positions 
of the different thermometers will stand as follows :— 
TABLE IX. 
Thermometers Depths in true De : 
pths in Terms of 
Hacc ae fe Conductive Equivalents. 
17 0 0 
II. 3 3 
ELT, 9 34+ togn x 6= 8:58 
"1128 
According to this way of reckoning depths, we have the following rates of varia-— 
tion of the logarithmic amplitudes, and of the epochs separately, reduced from — 
the previously stated means for the whole period of eighteen years :— f 
TABLE X. 












Rates of Diminution 
‘ of Logarithmic Am- 
Portions of Rock, plitude per French ° 
foot, and Conductive 
Equivalents, 



Rate of Retardation 
of Epoch per French 
foot, and Conductive 
Equivalents. 


Between Thermometers Nos. I. and II. 1286 
Pry or II. and IIT. "1265 
III. and LY. +1286 





Between Thermometers Nos. I. and IV. +1252 

28. Comparing this Table with the preceding Table VIII., we see that the dis- 
crepancies are very much diminished; and we cannot doubt but that the con-— 
ductive power of the rock is less in the lower parts of the rock, and that the 
amount of the variation is approximately represented by Table IX. We have, © 
however, in Table X. still too great discrepancies to allow us to consider variation 
in the value of ke, as the only appreciable deviation from Fourier’s conditions of” 
uniformity. 
29. In endeavouring to find whether these residual discrepancies are owing to 

