1903.] of the Spontaneous Radiation of Radium. 197 



Taking this view of the luminous radiations visible to the eye, it 

 seemed highly probable that the molecular motions by which they 

 were set up, whether we suppose all the radium molecules alike to be 

 concerned, or those only which are in active change, would be so far 

 analogous to the vibrations produced artificially, when radium vapour 

 is rendered luminous in a flame, or by the blow of an electric discharge, 

 as, in like manner, to set up radiations of certain definite wave-lengths 

 or, in other words, to furnish a spectrum of bright lines. 



A preliminary prismatic examination of the glow from pure radium 

 bromide was attempted by eye. In consequence of the feebleness of 

 the light under dispersion a slit spectroscope could not be used. A 

 thin fragment of some length of radium was selected, which in the 

 dark shone as a narrow line of light ; when this was viewed through 

 a direct-vision prism, it was seen to be dispersed into a spectrum which 

 extended from the blue down to about D where it became too faint to 

 be traced farther in the direction of the red. Within this faint 

 spectrum certain spots were distinctly brighter, due, in all probability, 

 to the presence of bright lines at those positions in the spectrum. 



The success of this preliminary observation encouraged us to hope 

 that it might be possible by availing ourselves of the accumulative 

 power of continuous photographic exposure, to obtain a record of the 

 blue, violet, and ultra-violet regions of the spectrum, if the glow radia- 

 tions extended so far. 



We made use of a small quartz spectroscope which had been con- 

 structed some years ago for very faint celestial objects. It consists of 

 a compound quartz prism of 60, consisting of two prisms of 30 of 

 right-handed and left-handed quartz respectively. The quartz lenses 

 are of short focus and of large angular aperture, being about ^ f. 

 The focal length of the lenses is 5| inches ; they are plano-convex, the 

 marginal parts of the convex surfaces being " figured " to diminish 

 spherical aberration. 



The solid radium bromide was placed at about a millimetre distance 

 in front of the slit, which had to be wider than if a bright object was 

 being photographed ; the width was about T i(yth inch. In the case of 

 the spark spectrum of radium and the comparison spectrum of nitroger, 

 a slit of less than half this width was used. 



With an exposure of 24 hours, faint traces of two lines were seen 

 on the plate. After several trials the negative reproduced on the 

 accompanying plate was obtained with an exposure of 72 hours. The 

 reproduction is enlarged two and a-half times. The spectrum consists 

 of eight bright lines, and at least eight faint lines, together with a 

 faint trace of continuous spectrum in the blue region, which does not 

 come out in the reproduction. 



It was seen at once that the two very strong characteristic rays of 

 the spark spectrum of radium, in this part of the spectrum, namely, 



P 2 



