On Underground Temperatures. 43 



fewer the changes of tubes, the differences in this respect, together 

 with the differences in length of the tube, may possibly help to 

 explain the reason why the wells of 300 to 600 feet deep gave gene- 

 rally a more rapid rate of increase than the wells of 2000 feet, and 

 may also account for the different rates of increase at the several 

 wells, for which otherwise there is no apparent cause. 



I have excluded a large number of overflowing wells because 

 of the uncertainty which attaches to the instruments used, or to some 

 essential point of which we are in ignorance, such, for example, 

 as the influx of other springs, the precise depth, the size of the 

 tubes, &c. These reasons apply to such wells as those of Newport 

 (No. 210), Falkirk and Midlothian (45), Dunkirk and Bourbourg 

 (209, 208), Alfort (49), Meaux (62), Arcachon (183), and others, 

 where we do not know whether or not standard and protected instru- 

 ments were used, or whether the experiments were in all such cases 

 made by competent observers. 



With respect to the extra- European observations, still greater un- 

 certainty attaches from our ignorance of the general conditions, and 

 especially of the exact mean annual temperature of the several places. 

 At the same time there are some exceptions worthy of consideration. 

 The experiments in the Sahara Desert (No. 88a) were made by an 

 engineer of great experience in the construction of artesian wells, and 

 accustomed to observations of this description, and are the mean of 

 results obtained at a number of wells. The observations at Charleston 

 and St. Louis appear reliable, only in these cases more particulars are 

 desirable. 



The African and Indian experiments seem to indicate a more rapid 

 rate of increase of temperature with depth than occurs in Europe. 

 Not so the American (U.S.) observations, which appear to indicate con- 

 ditions very similar to those which obtain here. Not much weight can 

 be attached to the solitary observation in South America. It requires 

 confirmation. 



IV. Tunnels. 



The few observations of this class, limited as they are, show not 

 only the modifications of the gradients caused by inequalities of 

 surface, but bear also on some important geological questions con- 

 nected with the structure of mountain chains and metamorphism. 

 The first great tunnel was that of Mont Cenis, which is about 7 miles 

 long, and passes under a ridge of the Alps rising 9532 feet above the 

 sea level, and 5280 feet above the tunnel. After making a correction 

 for the convexity of the surface, Professor Everett estimates the rate 

 of increase of temperature with depth to be 1° F. in 79 feet. But 

 the observations there were commenced late, and were not very 

 complete. 



