DR. \V. CKOKFKKY DUFFIELD ON THE 



In the present investigation quartz windows were discarded in favour of glass ones 

 for use in the pressure cylinder on account of the greater tendency of the former to 

 collapse after being suhjected to the highest pressures and the attendant expense of 

 replacing them. This, unfortunately, involved a restriction to the wave-lengths 



above 3550 A.U. 



The writer has not found any reference to previous work with the gold arc und< 



pressure. 



2 The Spectrum of the Gold Arc at Atmospheric Pressure. In these experiments 

 it was decided to employ the spectrum emitted by an arc passing between poles 

 consisting of rods of the pure metal. Previous workers with the gold arc at 

 atmospheric pressure have fed the metal into a carbon arc, but this method did not 

 seem desirable in the present experiments, in which it has been the writer's aim to 

 keep the element under investigation as free as possible from any sensible quantity of 

 a different element, which might by its own vapour density introduce some disturbing 

 effect upon the broadening or displacement of the lines. 



Thanks to the courtesy of Messrs. JOHNSON and MATTHEY, of Hatton Garden, 

 London, it was made possible to carry out these experiments with poles of metallic 

 gold. Each pole was of an inch in diameter and l| inches long, and was screwed 

 into a copper rod of the same diameter, 6 inches long, to enable it to be properly 

 placed in the pressure cylinder. The gold was guaranteed to be of purity 0'9995. 



The current employed was supplied by the Corporation mains at 100 volts 

 continuous, about 7 "5 amperes being taken by the arc at atmospheric pressure. 



The following Table I. gives a list of the lines which appear upon the photographs, 

 together with references to the observations made by previous workers in the same 

 part of the spectrum. 



It will be noticed that in the present photographs a number of lines appear which 

 were not described by KAYSER and RUNGE as occurring in the arc spectrum, though 

 most of them have been recorded by EDEE and VALENTA and EXNER and HASCHEK 

 as spark lines. The difference between KAYSEK and RUNGE'S and the writer's lists is 

 to be looked for in the use of metallic poles instead of a carbon arc fed with the 

 substance, and also perhaps in the current and voltage of the supply. 



The fact that the majority of the lines which appear in the arc under discussion 

 had previously only been found in the spark spectrum points to the insufficiency of 

 the terms " arc " and " spark " lines to distinguish lines from one another, and indicates 

 that in the case of gold, as in that of iron*, "spark lines" [sic] may appear in the 

 spectrum obtained from an arc source. The distinction implied by the terms " arc" 

 and "spark" lines cannot, therefore, be a general and rigorous one, however useful as 

 broad descriptive terms, but may have particular value when the current, dimensions 

 the poles, and voltage of the supply are known in the one case, and the induction, 

 capacity, &c., in the other. 



* DUFFIKLD, ' Astrophys. Journ.,' voL XXVII., p. 260, 1908. 



