On Underground Temperatures. 



39 



3. The first experiments at Grenelle were made when the bore-hole 

 had reached the depth of 400 metres (1312 feet), and when no work 

 had been going on for some weeks. The thermometers, of which three 

 sets were used, were left down thirty-six hours. They were protected 

 against pressure, and Arago remarks that the chalk through which 

 they were boring made so thick a paste filling the bore-hole, that 

 convection currents were hardly possible. All three sets of instru- 

 ments gave results within a fraction of a degree to one another.* 



4. In the bore-hole of the well at the Ecole Militaire the experi- 

 ments were made by Walferdin under very similar conditions. 



5. Pregny near Geneva. — The thermometer was here protected 

 against pressure, and to a certain extent against convection currents. 

 The bore-hole was small, and the water stood at a small depth below 

 the surface. 



6. The observations at the well at Ostend, with the water also 

 slightly overflowing, were made by Professor G. Dewalque, of Liege, 

 with protected thermometers at the bottom of the bore-hole. 



7. At Sperenberg, especial care was taken against convection cur- 

 rents, though how far these currents may have operated in reducing 

 generally the temperature in the lower part of the bore-hole, before 

 the experiments commenced, may be a little uncertain. 



These seven wells give a mean rate of increase of rather more than 

 51 feet for each degree F. in depth (see p. 40). The reason why I have 

 not included a larger number, is because in these cases I believe in 

 the existence of some undetected errors, such, for example, as, amongst 

 many others, those mentioned in the following instances. 



Swinderby. — I have already explained my objections in this case. 

 Of the first series of observations down to a depth of 1500 feet, Pro- 

 fessor Everett remarks " it is obvious that nearly all the tempera- 

 tures are largely affected by convection," he considered the bottom 

 temperature at 2000 feet as less likely to be vitiated by convection in 

 consequence of the small diameter of the bore-hole. 



[But if we take, as I have suggested, a mean depth and a mean 

 temperature, I believe we should have in the thermometric gradient of 

 47 feet per degree, a nearer approach to the true normal at Swinderby. 

 Even in cases where the temperature is uniform from top to bottom, 

 and there appears to be no clue to follow, as in the instance with the 

 Moscow Well (No. 135), where Professor Lubinoff records a tempera- 

 ture of 10-1° C. (50-2° F.), if not for the whole depth of 994 feet, at least 

 from 350 to 994 feet, it seems possible to obtain an approximate 

 gradient. For the mean annual surface temperature being 39"5° F., 

 and the half depth 494 feet, if we divide this by 10*7° (the difference 



* I have adopted this observation for Grrenelle, in preference to that on the over- 

 flow water, for the reasons given further on. 



