66o 



NATURE 



[August 28, 19 13 



which left a scarcely visible residue was intensely 

 radio-active, and from its solution a minute trace oi 

 a dark-coloured sulphide was precipitated by sulphur- 

 etted hydrogen. This latter was separated by centri- 

 fuging the solution in a capillary tube, dissolved in a 

 drop of nitric acid, and tested by the Behrens method. 

 The drop yielded a copious crop of triple nitrite 

 crystals, and the amount of lead present was estimated 

 at 1/4000 milligram. Now in the whole of the dis- 

 integration products in the tube only about 1/7000 

 milligram of radium-G was contained, and, further, 

 only a fraction of the total matter was used for the 

 test. Hence it seems probable that radium-G and 

 radium-D both form a characteristically crystalline 

 triple nitrite identical with that formed by lead. 



An attempt was made to determine what fraction 

 of the whole of the matter originally in the tube was 

 present in the drop tested, by comparing the 0-radia- 

 tion of the various portions, but no great trustworthi- 

 ness could be placed on the measurements at the time, 



found in the yellow line a satellite having anomalous 

 character. Wendt noticed four characteristic lines in 

 strong fields, but the satellites from which these lines 

 proceed could not be identified. 



Owing to extreme complexity in the distribution of 

 lines when the field strength is increased, we have 

 as yet no continuous observation indicating the posi- 

 tion of satellites of mercury lines in different mag- 

 netic fields. 



To fill in this gap, we made measurements on the 

 satellites of the violet line 4047 and the green line 

 5461, up to the field of 30 kilogauss, and arrived at 

 an unexpected result that the satellites show anomalous 

 Zeeman effect, either as regards intensity or the 

 mode of separation, and in nearly all cases both 

 combined. 



The instrument used was an echelon grating, but 

 it was sometimes crossed with a Fabry-Perot air plate 

 or a Lummer-Gehrcke plane-parallel plate to eliminate 

 the false lines. From a large number of photographs 







,u <->■ 



since the chloride of radium-E has probably a different 

 solubility to the chlorides of D and G. In a few 

 months' time, however, when the /3-radiation of E 

 has reached its equilibrium value a repetition of the 

 measurements should lead to a definite conclusion. 

 Hence, excluding the possibility of undetected sources 

 of error, one must conclude that the slow change 

 products of the disintegration of niton contain a body 

 behaving like lead, and, further, that this body is 

 either radium-D alone or a mixture of radium-D and 

 radium-G. Thus the investigation, so far as it has 

 gone at present, tends further to support the theories 

 of Soddy and of Fajans. R. Whyii.aw-Gray. 



University College, London. 



Anomalous Zeeman Effect in the Satellites of 

 Mercury Lines. 



Gehrcke and Baeyer, Lunelund, and McLennan 

 showed that the satellites of some mercury lines are 

 separated in a regular manner in weak fields; Gmelin 



NO. 2287, VOL. 91] 



of transverse effects obtained in different fields, which 

 were mostly uniform but sometimes heterogeneous, 

 the positions of the satellites were plotted and traced 

 into continuous curves ; these are given in the accom- 

 panying figures, for components vibrating parallel and 

 perpendicular to the field. The branches of the prin- 

 cipal lines are indicated by P, and shaded to show the 

 actual breadth; the positions of satellites in zero field 

 are marked on the axis of abscissa by lines proportional 

 to the intensity. 



A glance at the figures for parallel components 

 show that the separation is not generally proportional 

 to the magnetic field. The satellites of 4047 are 

 characterised by approaching the central principal line 

 asymptotically, while in we^V fields, the change in 

 wave-length is quite rapid. Tne.most remarkable of 

 them is the satellite at —60 m.A.LT. from the prin- 

 cipal ; the branch towards the positive side is approxi- 

 mately a parabola with vertical axis, and' that towards 

 the negative side a similar curve with horizontal axis ; 

 consequently the chang'e wrought by magnetic field is 



