420 Sir William Crookes .
[ Dec. 8 Another interesting point is raised by these results .
In free space the propagation of waves in a straight line is quite independent of any statical electric or magnetic field .
But if the waves are propagated through a part of space containing matter , the streaming of the ions produced by the waves seems to lead to the conclusion that the propagation of the waves is no longer independent of the statical electric and magnetic field , and aberration must result .
In conclusion , I wish to express my great obligation to Professor Gray , who has discussed these results with me and read the paper with very great care .
[ Note added January 30.\#151 ; Lord Kelvin has expressed the view that a. radio-active body may in some way extract energy from the aether and again radiate it .
Professor and Madame Curie have also suggested a possible abstraction from the surrounding gas .
The results obtained here support such views and indicate in some measure how such a process of selection may go on .
] " On the Ultra-Violet Spectrum of Gadolinium."* ' By Sir William Crookes , D.Sc .
, F.R.S. Received December 8 , \#151 ; Read December 15 , 1904 .
Gadolinium oxide is a rare earth , occurring between yttrium and samarium .
It was discovered in 1880 by Marignac , and was at first called by him Ya , a designation which he soon changed for gadolinium .
Since Marignac 's time much work has been done on this earth by Lecoq de Boisbaudran , Bettendorf , Cleve , Benedicks , Marc , Demar\lt ; jay , Exner and Haschek , Urbain , and others .
In the spring of this year , M. G. Urbain gave me some gadolinia and other rare earths , which he had prepared in a state of considerable purity by means of a novel system of fractionation in which use is made of the crystallisation of double nitrates of bismuth and magnesium with the rare earth nitrates .
He finds that bismuth places itself between the ceric and the terbic groups , thus sharply separating samarium , the last member of the ceric group , from europium and gadolinium , the first members of the terbic groups .
I have for some time past been fractionating rare earths by Urbain 's method , and can quite corroborate what he says .
The ultra-violet spectrum of gadolinium has been measured by Exner and Haschek , who have published their results in a book , [ * A plate of the spectrum to which this communication refers will , it is hoped , be published in another place .
] 1904 .
] On the Ultra-violet Spectrum of Gadolinium , 421 \#171 ; Wellenlangen-Tabellen fiir Spektralanalytische Untersuchungen , " * These observers used a material prepared by Dr. L. Haitinger , in which :they say holmium , samarium , and yttrium were present as impurities .
They give wave-lengths of 1150 lines ; I have found them to be as a rule very accurate , and in the maps accompanying this I have adopted most of their wave-lengths , except in cases where my own measurements give different figures , or where lines given by them are not to be distinguished on my photographs .
In some cases also lines ascribed to gadolinium appear to be caused by other earths as impurities .
My photographs were taken by means of the apparatus described in my paper " On the Ultra-Violet Spectrum of Radium .
" !
The arrangement is the same as there adopted in the maps of the radium spectrum .
The upper half of each strip shows the iron lines used as standards , with their wave-lengths according to Rowland 's latest measurements .
The lower half contains the gadolinium lines , with their wave-lengths i re-calculated or verified from the iron standards , according to the method given in detail in the paper just quoted .
Other lines , the wave-lengths of which are written in red ink , are either platinum lines or lines due to traces of impurity in the gadolinium oxide .
These are very few , and their faintness speaks well for the accuracy with which M. Urbain has separated other earths from the gadolinium , for in my experience I have seldom found a so-called " pure " salt of an ordinary metal anything like so free from impurities as this earth proves to be .
Moreover , it must not be taken as certain that some of the impurities assumed to be present on the strength of a strong line are really there , because in some instances I have not been able to detect the presence of another equally characteristic line of the same impurity .
The rare elements allied to gadolinium , or occurring with it , which I have thus found to be present , are ; Yttrium , represented by the lines at Europium " " " Samarium , , " " Ytterbium " " " Scandium , , , , , , 3774-51 and 4177-68 .
3972-16 , 4129-90 , and 4662-08 , 3230-65 .
3289-52 and 3694-35 .
3572-71 , 3613-96 , and 3630-86 .
The other elements which are present are\#151 ; Bismuth , represented by the line at . . .
... 4259 " 85 .
Magnesium " " lines at ... ... .
2796*62 and 2802-80 .
Calcium " " " ... ; 3933*81 and 3968-60 .
Of these , ' the bismuth and magnesium are present from the salts * F. Deuticke , Leipzig and Wien , 1902 .
f * Roy .
Soc. Proc. , ' vol. 72 , p. 295 .
, Aug. , 1903 .
422 On the Ultra-violet Spectrum of Gadolinium .
used in the fractionation , * while calcium is represented by the great H and K pair , which are almost always present in earthy spectra .
I have also examined the phosphorescent spectrum of Urbain 's gadolinium by means of photography .
In 1886 I communicated a paper to the Royal Society , t on the visible spectrum of phosphorescing Ya ( gadolinia ) , given to me by Marignac , and also prepared by myself , and later in the year J I said that the gadolinium I was then working with showed , on the evidence of its phosphorescent spectrum , the presence of samarium and other impurities .
Since that time I have used the photographic method of examining phosphorescent spectra , and in 1899 I brought before the Royal Society , S a preliminary notice of a new element , Yictorium , the existence of which I was led to infer from its chemical properties , and from some bands in its phosphorescent spectrum .
These bands consist of a strong group high up in the ultra-violet spectrum , having wave-lengths of 3117\#151 ; 3120 , .
3060 , 3064 , and 3219 .
The specimens of gadolinia prepared by myself in 1886 , as well as that sent me by Marignac , gave these victorium bands faintly , and the gadolinia given me by M. Urbain also showed- the same bands in greater strength than the other lines and bands I have ascribed to samarium and yttrium .
Early this month M. Urbain has sent me a. specimen of gadolinium which he considers quite pure , and on testing it in the vacuum tube for its phosphorescent spectrum I find the victorium bands photograph stronger than they came out in the other specimen .
At the same time I do not think that victorium is gadolinium .
All the evidence from chemical and phosphorescent data tends to the conclusion that victorium is only present as an impurity in the gadolinium , and the strength with which it reveals its presence is mainly due to the excessive delicacy of the test .
* In a still purer specimen from M. Urbain , recently received , these impurities , are absent .
+ ( Roy .
Soc. Proc. , ' vol. 40 , p. 236 , Feb. , 1886 .
$ ' Roy .
Soc. Proc. , ' vol. 40 , p. 503 , June , 1886 .
S * Roy .
Soc. Proc. , ' vol. 65 , p. 237 , May , 1899 .

