URANIUM AND GEOLOGY JOLY. 373 



general mean is (on my experiments) 7.1 million millionths of a gram 

 2)er gram. This mean is well distributed, as follows : 



Jurassic and Triassic altered sediments 6.4 



Crystalline schists, partly Jurassic and Triassic, partly Arch?eau_ 7.3 



Monte Leone gneiss and primitive gneiss 6.3 



Schistose gneiss (a fold from beneath) 6.5 



Antigorio gneiss 6.8 



The divisional arrangement is Professor Schardt's. Forty-nine 

 typical rocks are used in obtaining these results, and the experiments 

 have been in many cases repeated on duplicate specimens. Including 

 some ver}^ exceptional results, the means would rise to 9.1 X 10~^^ 

 grams per gram. 



Of the St. Gothard rocks I have examined 51 specimens, selected to 

 be, as far as attainable, representative.'^ 



Of these, 21 are from the central region, and their mean radium 

 content is just 3.3. The portion of the tunnel from which these rocks 

 come is closely coincident with Stapff's thermal subdivision of regions 

 of low temperature.^ This portion of the mountain offers the most 

 definite conditions for comparison with the Simplon results. The 

 region south of this is affected by water circulation ; the regions to the 

 north are affected hj the high temperature of the granite. 



"We see, then, that the most definite data at our disposal in compar- 

 ing the conditions as regards temperature and radio-thermal actions 

 in the two tunnels appear to show that the steeper gradient is asso- 

 ciated with the greater radium content. 



It is possible to arrive at an estimate of the downward extension of 

 the two rock masses (assumed to maintain to the same depth their 

 observed radio-activity), which would account for the difference in 

 gradient. In making this estimate, we do not assume that the entire 

 heat flow indicated by the gradients is due to radium, but that the 

 difference in radium content is responsible for the difference of heat 

 flow. If some of the heat is conducted from an interior source (of 

 whatever origin), we assume that this is alike in both cases. We also 

 assume the conductivities alike. 



Calculating on this basis, the depth required to establish on the 

 radium measurements the observed difference in gradients of the Cen- 

 tral St. Gothard and of the Simplon, Ave find the depth to be about 7 

 kilometers on the low mean of the Simplon rocks and 5 kilometers on 

 the high mean. There is, as I have already said, nothing improbable 

 in such a downward extension of primitive rocks having the radio- 



"I would like to express here my acknowledgments to the trustees of the 

 British Museum for granting me permission to use chips of the rocks in their 

 possession, and especially to Mr. Prior for his valuable assistance in selecting 

 the specimens. 



^ Trans. North of England Min. and Mech. Eng., Vol. XXXIII, p. 25. 



