278 Messrs. F. Soddy and T. D. Mackenzie on the 



details o£ the instrument, for it suffices to follow approxi- 

 mately the measurements given to obtain an instrument giving 

 without calibration a fairly accurate measure for excessively 

 minute quantities of radium. The instrument described 

 should certainly be trustworthy for absolute measurements 

 of minute quantities of radium without being calibrated to 

 within at most 50 per cent, over a range from, say, 10 ~ 9 to 

 10 ~ n gram of radium. 



Calibration of the Electroscope. — The relative sensitiveness 

 of the electroscope was determined, by measuring the leak 

 caused by the 7 rays from a constant amount of radium placed 

 in a definite position with regard to the instrument. The 

 radium employed wasO'2 gram of radiferous barium chloride 

 from the Societe Centrale de Produits Chimiques, of initial 

 activity a thousand times uranium, and at the time of writing 

 about five years old. It was placed, in its original unopened 

 sealed glass tube, within a lead tube of wall thickness 5 mm. 

 and the ends of the latter were closed by hammering. This 

 was placed 25 cm. below the base of the electroscope, 11*4 cm. 

 of this distance being pine-wood (the table and block on 

 which the instrument rested) and the remaining 13' 6 cm. 

 air. 



It seemed reasonable at the commencement of the work to 

 anticipate that the "standard 5 ' would slowly increase in 

 activity with lapse of time, owing to the generation of 

 Rutherford's Radium E, but the subsequent work of Meyer 

 and von Schweidler ( Wien. Anzeiqer, 12, Sitzunq. April 26th 

 1906) and Eve (Phil. Mag. 1906," [6] xi. p. 586) has shown 

 that no appreciable 7 radiation accompanies the 8 radiation 

 of Radium E. The most recent results of H. W. Schmidt 

 (Phys. Zeit. 1907, viii. p. 361) have established the existence 

 of such a 7 radiation, but it is so feeble, and relatively so low 

 in penetrative power, that it cannot affect the constancy of 

 the 7 radiation of the " standard " under the conditions 

 described. Contrary to the initial expectation therefore the 

 standard has probably not changed during the years the 

 investigation has been in progress. The average leaks pro- 

 duced by the standard are now somewhat higher than the 

 average two years ago, but it is impossible yet to say 

 whether the cause is in the standard or in the electroscopes. 

 In all, the sensitiveness of the electroscope has been determined 

 about twenty times in the last two years, and probably with 

 not less than ten new leaf-systems. The lowest reading re- 

 corded is 1*8, and the highest 3'2. The present mean is 

 about 2 "7, and this is taken as the mean sensitiveness of 

 an electroscope constructed as described. It is remarkable 



