334 
MR. T. ROYDS ON THE CONSTITUTION OF THE ELECTRIC SPARK. 
2. Aim of the Present Research .— Professor Schuster suggested that I should 
repeat and continue the experiments done by himself and Dr. Hemsalech. It was 
hoped to obtain photographs with greater dispersion, greater resolving power, and 
improved definition, which would remove the doubts expressed by other observers 
regarding the interpretation of the photographs, and would elucidate still further the 
constitution of the spark. 
3. Apparatus Employed .— The arrangement ol the apparatus was substantially the 
same as that used by Schuster and Hemsalech, except that two prisms were used 
instead of one. The prisms, which were kindly lent by Dr. B. S. Hutton for this 
research, had each an angle of 60 degrees and a base of 5 centims. The glass is a 
medium flint, of high transparency to the ultraviolet. The prisms were tested 
separately by observing the sodium D lines produced by a sodium bead in the Bunsen 
flame. The definition appeared to be perfect, and was not improved by allowing 
light to pass through only a narrow portion of the prism. With these prisms and 
the same lenses as were used in the previous work the extent of the spectrum from 
F to H is 2 centims. 
The optical system was adjusted as described in the paper referred to, and the 
prisms placed in the position of minimum deviation for about A 4200, the centre of 
the region to be photographed. The same method of focussing the spectrum on the 
photographic film has been adopted. The photographic film is carried on the same 
disc as used by the earlier experimenters, and can be rotated by a motor at speeds up 
to 160 revolutions per second. The speed of the disc is given by that of the motor 
read by a direct-reading speed-indicator by Elliot Bros., which has been tested and 
found accurate. 
The sparks were obtained between metallic electrodes from a condenser charged 
by means of a large Wimshurst machine capable of producing, without condensers, 
sparks 12 inches long. The best form for the electrodes is conical. Their surfaces 
were prepared with fine emery cloth, and finally polished by rubbing with wash- 
leather to remove small projections which would cause preliminary discharges before 
the condensers were fully charged, or which might affect the results on account of the 
distribution of the electric force near them. This preparation of the electrodes, 
however, probably introduces the calcium lines H and K into the spectrum. The 
condensers consist of plates of glass coated with tinfoil, connected in parallel with 
each other. The total capacity is 0'0306 mfd. They could be divided so as to 
obtain capacities of 0'0206 and 0'0103 mfd., roughly two-thirds and one-third of 
the total capacity. 
The improvement of the definition of the photographs reproduced in this paper 
over those of Schuster and Hemsalech is due to the improved optical quality of the 
prisms used and the greater care taken in preventing a backward or forward displace¬ 
ment of the rotating disc. 
4. Measurement of the Photographs. —If both electrodes are of the same material, 
