NORMAL GEOTHERMAL GRADIENT 113 
On account of the rapid diminution of the number of wells with 
depth (Table IV) the accuracy of the preceding sequences of numbers 
diminishes rapidly as the depth increases. 
_No attempt has been made in this paper to extend an average 
depth-temperature curve beyond a depth of 4,o00 feet. Such an 
extension probably involves the average thickness of the sediments 
and other factors that have not been investigated. 
CONCLUSIONS 
Practically nothing is known in regard to the positions of the 
isogeotherms over metalliferous deposits, such as are found in north- 
ern Ontario, Canada, and elsewhere throughout the world. As a 
result of ascending waters that have long since become extinct, it is 
reasonable to infer that the isogeotherms in many of these areas are 
depressed relative to those in the immediately surrounding zone. How- 
ever, granting this possibility, it seems rather improbable that the 
reciprocal gradients in the extended zone surrounding these areas 
of very low temperatures should be less than 100 feet per °F. (54.9 
meters per °C.); nor is it altogether improbable that the rate may 
not exceed 200 feet per °F. (109.7 meters per °C.). 
Comparison of this low rate of 1° F. in 200 feet in the slightly 
disturbed rocks of the Canadian shield with the preceding rates (49.8, 
62.1, 65.2) shows, as suggested on page 111, that there are at least two 
distinct types of normal gradient—one for sediments and one for 
exposed basement rocks that have remained static, or almost static, 
since their solidification. 
Apart from differences in the thermal constants of the rocks and 
the variation of the annual mean temperature of the air with elevation, 
latitude, and geologic climate, the average of even a comparatively 
small number of gradients in the undisturbed crystalline rocks must 
be nearly constant over the surface of the earth. 
In the sedimentary areas, however, uplift, subsidence, erosion, 
chemical reactions, and many other factors produce variations in the 
rock temperatures. Hence, a normal gradient determined from such 
areas depends on the number and distribution of the wells and the 
results of endless geological changes. For purposes of theoretical 
definition, the number of wells can be assumed to be so great that 
the average approaches a definite value, but, for practical purposes, 
it can not be assumed that the average obtained from a limited 
number of wells approaches a definite value. To illustrate, the recip- 
rocal of the mean gradient, from 1o wells in western Texas is, say, 
150; in Wyoming, again using 10 wells, the reciprocal is perhaps 40 
773 
