678 TRANSACTIONS OF SECTION C. 



iu Nature of these elements, uot only in association but in a quantitative relation- 

 ship, can only be explained on a genetic connection between the two. This 

 evidence, mainly due to the work of Boltwood, when examined in detail, becomes 

 overwhelmingly convincing. 



Thus it is to uranium that we look for the continuance of the supplies of 

 radium. In it we tind an all but eternal source. The fraction of this substance 

 which decays each year, or, rather, is transformed to a lower atomic weight, is 

 measured in tens of thousands of millionths ; so that the uranium of the earth one 

 hundred million years ago was hardly more than one per cent, greater in mass 

 than it is to-day. 



As radio-active investigations became more refined and extended, it was 

 discovered that radium was widely diffused over the earth. The emanation of it 

 was obtained from the atmosphere, from the soil, from caves. It was extracted 

 from well waters, lladium was found in brick-earths, and everywhere in rocks 

 containing the least trace of demonstrable uranium, and Rutherford calculated 

 that a quantity of radium so minute as -i'U x 10~^* grams per gram of the 

 earth's mass would compensate for all the heat now passing out through its 

 surface as determined by the average temperature gradients. In 1906 the Hon. 

 11. J. Strutt, to whom geology owes so much, not only here but in other lines of 

 advance, was able to announce, from a systematic examination of rocks and 

 minerals from various parts of the world, that the average quantity of radium 

 per gram was many times in e.xcess of what Rutherford estimated as adequate 

 lo account for terrestrial heat-loss. The only inference possible was that the 

 surface radium was uot an indication of what was distributed throughout tbe 

 mass of tue earth, and, as you all know, Strutt suggested a world deriving its 

 internal temperature from a radium jacket some 45 miles in thickness, the interior 

 being free from radium.' 



My own e.'tperimental work, begun in 1904, was laid aside till after Mr. 

 .Strutt's paper had appeared, and a valued correspondence with its distinguished 

 author was permitted to me. This Address will bo concerned with the applica- 

 tion of my results to questions of geological dynamics. 



Did time permit i would, indeed, like to dvvell for a little on the practical 

 aspect of measurements as yet so little used or understood ; for the difficulties to 

 be overcome arc considerable, and the precautions to be taken many. Ihe quantities 

 dealt wiih are aslouudingly minute, and to extract with completeness a total of a 

 few billionths of a cubic millimetre of the radio-active gas — the emanation — from 

 perhaps half a litre or more of a solution rich iu dissolved substances cannot be 

 regarded as an operation exempt from possibility of error ; and errors of deficiency 

 are accordingly Irequently met with. 



Special ditiiculties, too, arise when dealing with certain classes of rocks. For 

 in some rocks the radium is not uniformly diffused, but is concentrated in radio- 

 active substances. We are in these cases assailed with all the troables which beset 

 the assayer of gold wlio is at a loss to determine the average yield ofa rock wherein 

 tbe ore is sporadically' distributed. In the case of radium determinations this 

 difficulty may be sj much the more intensified as the isolated quantities involved 

 are the more minute and yet the more potent to affect the result of any one 

 experiment. There is here a source of discrepancy in successive e.xperiments 

 upon those rocks in which, from metamorphic or other actions, a segregation 

 ot the uranium has taken place. \Vith such rocks the divergences between 

 successive results are often considerable, and only by multiplying the number 

 of experiments can we hope to obtain fair indications of the average radio-activity. 

 It is noteworthy that these variations do not, so far as my observations extend, 

 present themselves when we deal with a recent marine sediment or with certain 

 unaltered deposits wherein there has been no readjustment of the original fine 

 stale of subdivision, and even distribution, which attended the precipitation of 

 the uranium in the process of sedimentation. 



I3ut the difficulties attending the estimation of radium in rocks and other 



' Proc. M.S., Ixxvii. p. 472, and bcxviii. p. 150. 



