July 24, 19 13] 



NATURE 



543 



light of mercury, published in the Philosophical Maga- 

 zine of 1800 (vol. xx., pp. 249-53). 



The first use of the mercury arc as a source of light 

 in polarimetry appears to have been made just ten 

 years ago by two German workers, Disch and Schon- 

 rock, working independently (Disch, Ann. Phys., 1903 

 (IV.), vol. xu\, 1155; Schonrock, Zeit. Vereins Deutsch. 

 Zuck. Ind., Tech. Part, 1903, vol liii., 652). Through 

 the personal kindness of Mr. Bastian, 1 was enabled 

 about three years later to make use of the same source 

 of light in what is still, perhaps, its most convenient 

 form. The glass Bastian lamp was designed to burn 

 with the coils of the arc in a horizontal plane, and 

 was arranged to light automatically in this position. 

 It was with great delight, therefore, that I discovered 

 that, in spite of all warnings to the contrary, the 

 lamp would continue to burn for any length of time 

 with the coils raised into a vertical plane; in this 

 position one of the straight portions of the arc could 

 be focussed by a condenser directly on to the slit of a 

 spectroscope, and so used to illuminate the field of a 

 polarimeter. The lamp consumed very little current, 

 and could be connected directly to the ordinary light- 

 ing circuits without any risk of "blowing" the fuses; 

 it was cheap to purchase, and as the resistances 

 formed part of the holder of the lamp there was no 

 need for any auxiliary apparatus whatever. In view 

 of its special suitability for polarimetric work, it is to 

 me personally a matter of some regret that this 

 pioneer lamp has been displaced completely by the 

 more powerful arcs, encased in refractory silica glass, 

 which now adorn the exteriors of so many places of 

 amusement. . 



When using the mercury arc as a source of violet 

 light, account must be taken of the greatly reduced 

 sensitiveness of the eye to light of such short wave- 

 length. It is here that the silica mercury lamp has 

 proved of such great utility. I am indebted both to 

 Mr. Lacell, of the Silica Syndicate, and to the Brush 

 Electrical Engineering Company for allowing me, for 

 experimental purposes, to distort their well-considered 

 designs for commercial mercury arc lamps. Here, 

 for instance, is a horizontal lamp which has been 

 altered so that the arc can be seen at its greatest 

 intensity in an end-on position. At first the light was 

 liable to be obscured by globules of condensed mer- 

 cury. But by recessing the window it was kept 

 sufficiently hot to prevent condensation, and this diffi- 

 culty was effectively overcome. Even then, however, 

 the arc was not so convenient as one arranged in a 

 vertical plane, like the upturned Bastian lamp. It 

 was at this stage that I persuaded the Brush Com- 

 pany to modify for me their "Quartzlite" lamp by 

 twisting one of the terminal U-tubes into such a posi- 

 tion that it did not empty itself when the lamp was 

 raised into a vertical plane. The " end-on " lamp 

 and the vertical " Quartzlite " lamp have been de- 

 scribed in the Transactions of the Faraday Society 

 (1912, vol. vii., pp. 267-70), and were exhibited at 

 the Optical Convention of June, 1012. The lamp 

 shown in Fig. 1 has not been described previously. 

 It combines the merits of both of the preceding pat- 

 terns, and ran be used either horizontally or vertic- 

 ally, and either in a side-on or in an end-on position. 



The -'Pinch Effect." 

 One feature of the silica mercury-lamps is suffi- 

 ciently remarkable to deserve attention. When the 

 arc is" first struck bv tilting the lamp it fills the whole 

 of the bore of the half-inch tube which encloses it; 

 but, in accordance with Faraday's observation that 

 currents travelling in the same direction attract one 

 another, the parallel threads of current are drawn 

 together until finally, as you see, the arc is "pinched" 

 NO. 2282, VOL. 91] 



together into a thread occupying only about one-third 

 of the diameter of the tube. This pinching together 

 of the arc contributes substantially to its efficiency as 

 an illuminant in polarimetric and spectroscopic work ; 

 but it is not a suitable form for projection, which 

 demands, as a condition for successful work, a power- 

 ful point-source of light. _ 



If the current in the mercury arc is increased, the 

 pinching effect may extend to the point of breaking 

 the threads of current completely and so extinguish 

 the arc. 



It may be of interest to 



refer here to the well-known 



fact that the "pinch effect," 



which I have exhibited on a 



small scale in a mercury- 

 lamp, is of great importance 



in the electrical melting of 



steel on a large commercial 



scale. In that case a current 



of great magnitude, flowing 



through a mass of molten 



steel enclosed in a circular 



channel, sometimes causes 



the metal to pinch together 



to such an extent that the 



circuit is actually broken. 



The "pinching" apart and 



running together of the mass 



of molten metal produce a 

 somewhat thrilling display. 



The Mercury Spectrum. 



The mercury arc differs 

 from the carbon arc in giving 

 an extremely simple line- 

 spectrum, the chief features 

 of which are a yellow, a 

 green, and a violet com- 

 ponent. The yellow com- 

 ponent contains two lines, 

 separated by about twenty 

 units of wave-length as 

 compared with six units ' ' Li _ 

 for the yellow sodium side-on or end-on position. 



doublet; it shows up well 



in the spectrum, but on account of its duplex 

 character it is not suitable for use in exact measure- 

 ments. 



By means of powerful high-resolution apparatus, 

 such as the echelon spectroscope, the green line of 

 the mercury spectrum has also been shown to be com- 

 plex (Fig. 2) ; but in this case the components are so 

 close together that they do not in any way reduce the 

 value of the line as a source of monochromatic light. 

 The extreme brilliancy of this green line, its high 

 spectroscopic purity, and the ease with which it canbe 

 produced, have given to it an unrivalled position 

 amongst the various sources of monochromatic light 

 which are now available for polarimetric work. I can 

 say with confidence that no one who has worked with 

 the mercury-lamp will ever wish to return to the 

 sodium flame, which it is rapidly displacing both in 

 scientific and in technical laboratories. 



Dr. Gladstone directed special attention to the 

 strength of the violet lines in the spectrum, of one of 

 which he said that "this ray is situated far beyond 

 what is ordinarily considered the limit of the luminous 

 spectrum." This deep-violet component contains two 

 lines which are clearly visible in the spectroscope; 

 but they lie so near to 'the limit of visibility that their 

 presence can be shown most clearly with the help of 

 a fluorescent screen. The bright violet line is, from 

 the scientific point of view, one of the most valuable 



