140 Rev. 0. Fisher on the Residual Efect of a Former 



2b f L 

 But the difference —-=. \ e'^dfi varies as b; and if b be 



larger than 20°, which we have hitherto on good grounds 

 assumed to be its value on the hypothesis of an ice-sheet, 

 and the present surface at 52°, this difference will be propor- 

 tionately larger, and the mean gradient consequently smaller. 

 This might happen if the ground was not covered with ice, 

 and the temperature of the surface very low, as in Siberia at 

 the present day. A low gradient may therefore possibly 

 point to such a condition of the surface during a former cold 

 period, but our equations will not so strictly apply to that 

 case, because the warming of the surface would not have come 

 on so rapidly as on the hypothesis of a melting icy covering, 

 and b would be a function of the time. The remarkably low 

 gradient of l/223"7 has lately been observed by Prof. A. 

 Agassiz in a deep mine in the neighbourhood of Lake 

 Superior*. If we apply the suggestion just made to this 

 case, taking the gradient during the cold period at 1/51 and 

 the time since elapsed as 40,000 years, we find the former 

 temperature of the surface as low as — 66° Fahr. This is 

 inconceivable. And if, instead of an original gradient of 1/51, 

 we assume one of 1 60, and the time elapsed 34,013 years 

 instead of 40,000, making use of the result already obtained 

 for Wheeling, we obtain for the temperature during the cold 

 period — 30°, which is also incredibly low. But the discre- 

 pancy of these two results is an instance of what a large 

 difference a comparatively small change in our hypotheses will 

 occasion. 



It does not appear probable, therefore, that we can account 

 for so low a gradient as is stated to have been observed at 

 this mine by calling to our aid the residual effects of a former 

 cooling of the surface. We must wait for further information 

 about this case. 



On the whole, it seems that the question proposed, as to 

 whether there is any prospect of obtaining a date for the 

 glacial epoch from underground temperature, must be answered 

 in the negative. But Ave have found the character of the 

 traces which a former glaciation might be expected to leave 

 behind it, the principal one being simply a reduction of the 

 gradient. This would to some extent account for the differetit 

 gradients which have been found at different localities. 



If we assume the rise of temperature b, and could rely upon 

 two observations at known depths sc 1 and x 2 , we should then, 

 * Nature, vol. liii. p. 161. 



