TRAKSACJTIONS Ot' SECTION B. (j7o 



'there fire two statements here which, according to my experibiice, appear to 

 require modification. That potassium and sodium are more abundantly dis- 

 tributed than lithium is true, but that these are more widely distributed is not 

 strictly correct; nor can it be accepted as unquestionable that lithium is an 

 unlikely constituent of dust, glass, copper, &c. Evidence to the cootrary is 

 based upon facts divided into three categories— first, those derived from the 

 qualitative spectroscopic analysis of common ores and minerals usually associated 

 with the alkali metals ; secondly, analysis of the crude salts of the alkalis, such 

 as the Stassfurth minerals and nitrates from Chile and Bengal, show that they 

 contain lithium and rubidium, with not unfrequently caesium. Facts belonging to 

 the third category are derived from experimental evidence, which is both quanti- 

 tative and spectrographic, the source of the spectra being the oxybydrogen flame. 

 When half a gramme of material yields a photograph of the spectrum of litliium 

 on which the four chief lines are visible — namely, \X 6708, 460-J-07, 4132 93, 

 and 323282 — there cannot be less than 0'0Grf9 gramme of lithium present. 

 When only the lines 6708'0 and 4603-07 are visible, the quantity is not less or 

 more than O'OOil gramme. 



When only the red line is photographed the quantity is not more than 

 0'002 gramme, and with half this quantity the line ceases to be photographed. 

 It follows, therefore, that from the evidence afforded by the number of plates on 

 which this line appears there could scarcely be less lithium in the O'o gramme of 

 material analysed than 0'2 per cent. 



Further results have been obtained with several other metallic compounds, 

 but the sensitiveness of the flame reaction varies extraordinarily with the spectra 

 of different elements. 



Mr. Ramage and I found in a hundred and seventy common ores and minerals 

 potassium and sodium, and with these common elements rubidium and lithium 

 were very generally associated. Thus, of sixty-two iron ores, rubidium was found 

 in sixty-one. In sixteen red haematites, massive minerals of the purest type, 

 rubidium was contained in four. Where potassium and rubidium occurred lithium 

 was invariably found. It was found in limestones, in dust, in the Bessemer flame, 

 in ordinary pipeclay, tobacco pipes, and a great variety of siliceous minerals, 

 such as the Dublin granites ; in Donegal kyanite, which contains 98 per cent, of 

 aluminium silicate ; and in asbestos. It was found in dust which fell from the 

 clouds, in volcanic dust, in soot, in flue-dust from chemical works, and in that 

 from copper smelting and refining works. This last material contained lithium, 

 sodium, potassium, rubidium, and ciesium ; copper, silver calcium, strontium, 

 aluminium, gallium, indium, thallium, iron, nickel, cobalt, manganese, chromium, 

 lead, zinc, cadmium, and tin. Upon such evidence as this it is impossible to 

 corroborate the statement that potassium is a more widely distributed element 

 than lithium, or that lithium is an unlikely constituent of dust, glass, copper, &c. 



3, Thfi Liquefaction of Helium. By Dr. Kammerlingii Onnep. 



d. Anticipations and Experiments on the Liquefaction of Helium. 

 By Sir Jambs Dewar, F.K.S. 



5, A Demonstration on the Rapid Electroanalytical Separation of Metals. 



By Dr. Henry J. S. Sand. 



In the process of electroanalysis to be shovra the principle of very vigorous 

 stirring of the electrolyte has been combined with that of keeping the potential 

 of the cathode under control by means of an auxiliary electrode. It has thus 

 been possible very largely to extend the scope of electroanalytical methods. The 



1908. ' X s 



