262 CARNEGIE INSTITUTION OF WASHINGTON. 



always a presumption that the gradients observed in mines are ab- 

 normal or abnormally varied. Artesian wells are made in order to 

 utilize the subterranean circulation of water, and that circulation 

 involves the convection of heat, whereby the normal gradient is 

 necessarily disturbed. Oil-wells and gas-wells can be successful 

 only in regions where the strata encountered at different depths are 

 of diverse character, and the temperature gradient theoretically 

 changes in passing from rock of one character to rock of another. 

 The successful wells have their normal temperatures disturbed by 

 the expansion of gas ; the unsuccessful usually penetrate zones of 

 water circulation. As these three categories include practically all 

 the deep openings which have been made in the earth, it is evident 

 that the combination of their data yields no trustworthy index of the 

 normal downward increase of temperature. The arithmetic mean 

 of all their results has less authority than a single determination 

 made under proper conditions. The ideal determination is to be 

 obtained by boring in homogeneous rock not recently subject to 

 disturbances calculated to modify its heat distribution. And such 

 rock will not be exploited in intelligent search for any economic 

 material. The determination which shall be of service to the stu 

 dent of geophysics must be made by a boring planned and executed 

 for the special purpose. 



Special emphasis may be given to the fact that all deep mines and 

 all deep borings heretofore made have penetrated varying instead 

 of uniform material. They have, therefore, presumptively encoun- 

 tered changes in temperature gradient arising from differences in 

 material, and as it is not practicable to separate such variations of 

 gradient from the variations dependent merely on depth, the latter 

 variations can not be deduced from records now in existence. They 

 can be afforded only by a deep boring in homogeneous material. 



CONDITIONS TO BE SATISFIED IN THE SELECTION OF A SITE FOR 



A BORING. 



Uniformity of Rock Character. — The temperature gradient within 

 the earth's crust, or the temperature change per unit of vertical 

 distance, varies locally with the conductivity of the material (more 

 strictly, with the diffusivity, which is a function of the conductivity 

 and the specific heat). It may be subject also to other variation, 

 but the discussion of other sources of variation is practically impos- 

 sible if their effects are complicated with those arising from diver- 

 sity of rock character. It is conceivable that the thermal record of 



