8 



Prof. J. Prestwich. 



the measure of the increase of temperature with depth, should be 

 placed at that point below the surface, at which the changes of 

 annual temperature cease to have effect. Any such determination* 

 might also with advantage be applied to the correction of the past 

 observations. 



In the absence of this information, we have for the present to be 

 satisfied with the mean surface temperature, where known, and take 

 that as the datum line from which to calculate the descending rate. 

 Sometimes this has been the mode adopted, and the gradient cal- 

 culated from the surface ; at other times it has been calculated by 

 supposing the mean invariable surface -temperature to lie at a depth 

 of 50 feet, and taking that as the datum level from which to start. 

 Thus, some of the original estimates have been founded on calcula- 

 tions commencing from the surface, and others at 50 feet beneath it. 



Either, therefore, some of these estimates are too high in conse- 

 quence of the deduction of 50 feet from the depth, or others are too 

 low in consequence of not making an allowance for the zone of 

 variable temperature. But as the mean annual temperature of the 

 ground at the surface generally in these latitudes exceeds that of the 

 air by about 1°, it follows that if we take the mean temperature of 

 the air at the place of observation, and calculate the rate of increase 

 from the surface, instead of allowing 1° for the higher temperature? 

 of the ground and with a datum line 50 feet lower, we shall come 

 to nearly the same result. At the same time it must be admitted 

 that in so doing, we may sometimes be liable, especially with stations 

 of moderate depth, to make the rate of increase with depth less than 

 it should be. It is therefore the more important in future observations 

 to determine if possible the temperature at a depth of 50 to 60 feet — 

 or where the first bar of uniform temperature may happen to be — and 

 to calculate from that point the rate of increase of temperature 

 with depth. Failing this, starting from the surface with the mean 

 annual temperature of the place will no doubt give an approximate 

 result. Thus the mean temperature of Paris is 51° F., and the tem- 

 perature of the cellars of the Observatory at a depth of 95 feet is 

 53° F., or 2° higher, so that if in this instance, whether we start with 

 a surface temperature of 51° or a temperature of 52° at a depth of 

 50 feet, the result would be almost exactly the same. 



In looking over Table I, the essential differences in the results ob- 

 tained in Mines, Coal-pits, and Artesian wells will be at once apparent. 

 This arises both from the circumstance that not only in each of these 



* This was done in the case of the experiments at Grrenelle, where a datum line 

 based on the temperature and depth of the cellars of the Paris Observatory was 

 employed. Tbese are 29 metres (95 feet) deep, and the thermometer stands in- 

 variably at 11-7° C. (53° F.), the mean annual temperature of Paris being 10*6° C. 

 (51° F.). This is, however, a greater depth than necessary. 



