1] RADIO-ACTIVE SUBSTANCES 17 
Later observations on the spectrum of radium have been made by 
Runge!, Exner and Haschek?, with specimens of radium prepared by 
Giesel. It has already been mentioned that the bromide of radium 
gives a characteristic pure carmine-red coloration to the Bunsen 
flame. The flame spectrum shows two broad bright bands in the 
orange-red, not observed in Demargay’s spectrum. In addition 
there is a line in the blue-green and two feeble lines in the violet. 
14. Atomic weight of radium. Mme Curie has made 
successive determinations of the atomic weight of the new element 
with specimens of steadily increasing purity. In the first obser- 
vation the radium was largely mixed with barium, and the atomic 
weight obtained was the same as that of barium, 1375. In 
successive observations with specimens of increasing purity the 
atomic weights of the mixture were 146 and 175, The final value 
obtained recently was 225, which may be taken as the atomic 
weight of radium on the assumption that it is divalent. 
In these experiments about 0:1 gram of pure radium chloride has 
been obtained by successive fractionations. The difficulty involved 
In preparing a quantity of pure radium chloride large enough to 
test the atomic weight may be gauged from the fact that only a 
few centigrams of fairly pure radium, or a few decigrams of less 
concentrated material, are obtained from treatment of about 2 tons 
of the mineral from which it is derived. 
Runge and Precht*? have examined the spectrum of radium in 
a magnetic field, and have shown the existence of series analogous 
to those observed for calcium, barium, and strontium. These series 
are connected with the atomic weights of the elements in question, 
and Runge and Precht have calculated by these means that the 
atomic weight of radium should be 258—a number considerably 
greater than the number 225 obtained by Mme Curie by means of 
chemical analysis. Marshall Watts‘, on the other hand, using another 
relation between the lines of the spectrum’, deduced the value 
obtained by Mme Curie. Considering that the number found 
1 Astrophys. Journal, p.1, 1900. Drude’s Annal. No. 10, p. 407, 1903. 
* Sitz. Ak. Wiss. Wien, July 4, 1901. 3 Phil. Mag. April, 1903. 
4 Phil. Mag. July, 1903. 
° Runge (Phil. Mag. Dec. 1903) has criticised the method of deduction em- 
ployed by Marshall Watts on the ground that the lines used for comparison in the 
different spectra were not homologous. 
Ry Rak; 2 
