272 REPORTS ON THE STATE OF SCIENCE, ETC. 



place to place is therefore of fundamental geophysical interest. To estimate 

 this heat flow it is necessary to know the vertical temperature gradient and 

 the thermal conductivity of the rocks in which the gradient is measured. 

 There exist numerous measurements of temperature in deep bores in various 

 parts of the world, but almost no conductivity data except that collected by 

 the former British Association Committee about fifty years ago. Thus 

 there is no trustworthy data on the variation of heat flow from place to 

 place, though it is believed by many that considerable variations occur. 



In an attempt to remedy this state of affairs the Committee has pursued 

 investigations along the following lines : 



(i) An attempt has been made to get the necessary data from shallow 

 holes. This investigation has met with difficulties through dis- 

 turbances of the temperature by percolating water. 



(2) Temperature measurements have been made in bores whenever they 

 became available. 



(3) An apparatus has been constructed for the measurement of thermal 

 conductivities of rock specimens. 



In the past it has been somewhat optimistically assumed that the con- 

 ductivity measured in the laboratory was the proper quantity to use in the 

 heat flow calculations. As the temperature gradient in the laboratory is 

 of the order of 10° c./cm. and that in nature 0-0003° c./km., this seems a 

 somewhat unsafe assumption. The investigations on heaters in shallow 

 holes and on the annual temperature wave employ gradients of the order 

 of 0-03° c./cm. Comparisons of these with the laboratory determinations 

 therefore provide a most valuable check. 



2. Measurements in a Shallow Hole. — If the temperature distribution is 

 steady the flow of heat per cm^ of the earth's surface should be independent 

 of depth. It would therefore be supposed that the heat flow could be 

 measured as well in a shallow hole as in a deep one, so long as the hole 

 was deep enough to get below the eflfect of the annual temperature wave. 

 This would avoid the troubles associated with the use of deep bores (see § 3) 

 and would have the added advantage that the conductivity could be 

 measured in situ by the temperature distribution round buried heaters. 



Experiments briefly described in last year's report showed that tempera- 

 tures could be measured with thermocouples in a shallow hole with an 

 adequate accuracy. A heater was installed in a is-ft. hole in gault. When 

 it was turned on the temperature of the thermo-junctions changed in the 

 expected way. Examples are shown in Fig. i. The change due to the 

 annual temperature wave was subtracted, and expressions of the form 

 A{i — Erf B/t) fitted to the results. The constants A and B are functions 

 of the conductivity, the specific heat per unit volume and of the positions 

 of the thermo-junctions. The conductivity deduced from them and from 

 laboratory measurements of the specific heat was 0-0027. The heat flow 

 could not be deduced from these measurements since there was a large 

 annual temperature change even at the bottom of the hole. The results, 

 however, were taken as indicating that the method was sufficiently promising 

 to try in a deeper hole. A loo-ft. hole was therefore drilled at a cost of 

 £19 in a field near the Observatory at Cambridge. Three feet of water- 

 bearing gravel were encountered on top of the gault and the top 20 ft. of 

 the bore was cased to exclude this water. In spite of this water continued to 

 enter the bore from lower levels. The casing was therefore continued to 

 60 ft., still without stopping the water ; the water level was different inside' 

 and outside the casing, showing that the water was really derived from the 

 gault and not from the surface gravel. As the hole showed signs of caving 



