258 REPORTS ON THE STATE OF SCIENCE, ETC. 



THERMAL CONDUCTIVITIES OF ROCKS. 



Report of Committee appointed to investigate the direct determination of the 

 Thermal Conductivities of Rocks in mines or borings where the tempera- 

 ture gradient has been, or is likely to be, measured (Dr. Ezer Griffiths, 

 F.R.S., Chairman ; Dr. E. C. Bullard, Dr. H. Jeffreys, F.R.S., 

 Dr. E. M. Anderson, Prof. W. G. Fearnsides, F.R.S., Prof. G. 

 Hickling, Prof. A. Holmes, Dr. D. W. Phillips, Prof. J. H. J. 

 Poole). 



Note on Radioactivities of Igneous Rocks. 



By Harold Jeffreys, F.R.S. 



Radioactivities of rocks of the same type are far from uniform. The avail- 

 able determinations have been rediscussed in the hope of improving 

 estimates of the mean radioactivities of the crustal layers and obtaining 

 criteria of their accuracy. 1 The general increase of Ra and Th with silica 

 content has been confirmed, but at the same time the variability increases, 

 not only absolutely, but in comparison with the mean. Only the plateau 

 and Pacific basalts show such an approach to uniformity as would entitle 

 us to infer that they have any resemblance to a uniform parent rock. For 

 rocks of the same type from different regions the means vary by much more 

 than can be attributed to random sampling, but the ratio of the standard 

 (mean square) departure from the mean to the mean itself is as nearly 

 constant as we could expect. It appears therefore that this variability 

 relative to the mean can be regarded as a property of the rock type. Like 

 the mean it increases for the sequence dunite — plateau basalt — basalt — 

 granite ; though the agreement between the dunites may be accidental. 

 The frequencies agree closely with the hypothesis that the chance of a 

 radioactivity in a range dx for a rock of given type for a given region is 



proportional to 



xP e {P + i)*lb dx 



where p is a constant for the type, but b varies with the region. For granites 

 p = 2-6, basalts, etc., 5-0, plateau and Pacific basalts, 30. The mean for 

 the region gives the best estimate of b. The following table gives some 

 estimated means with their standard errors, which are to be regarded as 

 minima, as some of the results are got by combining regions that may 

 turn out to differ systematically when more data are ready. The units are 

 io" 12 g/g for Ra, io _ 5 g/g for Th. 



Granites. 



Finland: Ra 4-66 ±0-40; Th 2-80 ±0-24. 



Alps : Ra 4-43 ± o-68 ; Th 3-30 ±0-50. 



Scotland, Ireland, N. America : Rai-59 ±0-12; Tho-8i ± 0-08. 

 Granodiorites. 



California : Ra 1 -77 ± 0-49 ; Th 2-35 ± 0-45. 

 Gneisses. 



Alps : Ra 3-26 ± 0-28 ; Th 1-75 ± 0-25. 



1 Gerlands Beitrage zur Geophysik, vol. 47, 1936, pp, 149-170. 



