192 



MI-SSRS. A. SCHUSTER AND G. HEMSALECH ON 



of the photographs was considerably increased when a much more powerful battery 

 was introduced, and charged from an induction machine constructed for us by 

 Mr. H. C. WIMSHUKST. This machine has 12 plates, of 62 centims. diameter, and 

 gives, without jars, sparks which are 13 inches long. In the majority of our experi- 

 ments six large jars of flint glass were used, each jar having a coating of tinfoil, of 

 about 2000 sq. centims. surface. The capacity of each jar was measured separately 

 for us by Mr. J. R. BEATTIE, and found to vary from '0049 to '0060 microfarad, the 

 capacity of the complete battery being '033 microfarad. The jars, which had not 

 been used for years, were found to be in bad condition. They were carefully 

 cleaned, by soaking first in caustic soda and then in nitric acid. After coating with 

 tinfoil to within 10 centims. from the top, the uncovered parts were heated in front 

 of a fire to over 60 C., so as to secure complete dryness, and then, whilst hot, 

 varnished. After this treatment the jars gave no trouble. 



Great care is necessary in handling a battery of this capacity, especially as the 

 experiments are conducted in the dark. 



Fig. 7. 



Fig. 6. 

 A 



1 



The whole battery was placed together on an insulated table, and Mr. WIMSHTJRST'S 

 recommendation to keep one hand always in the pocket while handling the jars also 

 proved useful. 



The metal electrodes were fitted inside a box having an opening towards the 

 collimator of the spectroscope. Perhaps the greatest source of trouble at first con- 

 sisted in the uncertainty of the position of the points on the electrodes, from which 

 the sparks set out. Successive sparks did not always leave the poles at the same 

 points, so that the spark images did not always fall centrally on the slit. This was 

 remedied to a great extent by giving careful attention to the shape of the electrodes 

 and to their polish. The form adopted after a few trials was conical, and is shown 

 in fig. G. The metal having first taen turned in the lathe to the required shape, was 



