On Underground Temperatures. 53 



The thermic curves attain their maximum variation in talcose and 

 micaceous schists and in slates. The greatest inequality, 1 : 3, was 

 shown by a specimen of a talcose rock, of sp. gr. 2*7. The variation 

 exists in all rocks showing schistosity or lamination, but in ordinary 

 stratified rocks the thermic curve remains that of the circle. It 

 was found that the variation exists also in rock crystal, gypsum, 

 felspar, &c. All the specimens experimented upon were dry. 



It is evident then that in gneissic rocks and slates, the dip, cleavage, 

 and foliation may have a very important effect on the conduction of 

 heat ; lamination has a similar but lesser effect in argillaceous shales ; 

 in ordinary sandstones and limestones no such effects are produced. 

 Whilst these effects therefore may be very manifest in the rocks 

 generally associated with Mineral Veins, they can only be small in 

 Coal Mines, although they may be in some places increased by a larger 

 proportion of mica in the sandstones and shales. There is also the 

 further consideration with strata, such as those of the Coal-measures, 

 that although there may be separately little difference in the 

 thermic axes of the different rocks, the differences of conductivity in 

 the various component strata may, as with foliation, allow of a 

 variable transmission of heat along the planes of the inclined 

 strata at their outcrop. But, even if that be the case, the effect would 

 be merely local, possibly affecting the mass of inclined strata to a 

 given depth, but in no ways affecting the special problem in relation 

 to the general body of strata unaffected by those local conditions. 



Conclusions. 



The list of selected cases on which these conclusions are based may 

 appear small, but I feel satisfied that the sources of error in experi- 

 ments of underground temperatures are so many and so obscure, 

 that without the fuller information which we have in these few 

 instances, the larger number are not available for our purpose, though 

 they all bear on the general question, and with the corrected data 

 before named it may be possible to utilise some of them hereafter. 

 We now require, however, for this object those observations only 

 which give the nearest possible approximation to the true thermo- 

 metric gradient, and for this it is necessary to reject all the doubtful 

 and more uncertain cases. For these reasons I have confined myself, 

 in the case of the Coal Mines, to the limited number of the eight 

 instances given in the list at p. 24; and in the Mineral Mines to 

 eighteen of the seemingly most reliable rock and spring observa- 

 tions of Fox and Henwood. The Artesian Wells give more uniform 

 results, — results which, under certain conditions, should be perfectly 

 true. I have, however, only been able to select fifteen wells, of which 

 eight are overflowing wells, and seven not-overflowing. 



