78 Messrs. Trowbridge and Richards on 



gators worked with voltaic cells which were not constant and 

 which required great watchfulness and continual renewals. 

 In onr investigations we are using a lead accumulator of the 

 Plante type, and rind it highly advantageous for spectroscopic 

 work ; for by means of the steady current afforded by it one 

 can study the spectra of gases under especially good con- 

 ditions. 



Our battery consists of five thousand cells, so arranged 

 that they may be disconnected and wholly reconnected in any 

 desired manner in less than a minute. The electromotive 

 force of the complete series is somewhat over ten thousand 

 volts, but when the cells are connected for quantity they may 

 be readily charged by means of a dynamo giving a tension of 

 only sixty volts. The insulation of the terminals of this 

 battery was a matter of some difficulty, for even dry wood 

 allows considerable leakage from one case of cells to another ; 

 but by the plentiful use of paraffin, mica, and vulcanite the 

 problem was solved with reasonable success. The discharge 

 from only a very small fraction of the battery produced a 

 most uncomfortable shock, and it is probable that the dis- 

 charge of the whole battery would be instantly fatal. The 

 great heat of this full discharge immediately shatters a 

 Geissler tube, the glass being splintered throughout the whole 

 length of the capillary. Hence a resistance of several million 

 ohms was usually interposed between the battery and the rest 

 of the apparatus, This resistance was also of service in pro- 

 tecting the experimenters from serious accidental shocks. 

 Ordinary distilled water contained in long tubes with movable 

 electrodes was the most convenient resistance for our purpose; 

 dilute solutions of cadmic iodide in amyl alcohol and of cadmic 

 sulphate in water between cadmium electrodes were also some- 

 times used. Unless these liquids are contained in tubes of 

 rather large diameter they are likely to cause irregularities 

 by boiling nnder the influence of the heat of the current. 

 Graphite resistances are too combustible to answer the purpose. 



The argon used in onr experiments was very kindly given 

 to one of us by Lord Rayieigh. It was a portion of the 

 purest preparation which had been used in his final determi- 

 nations of the density of the gas ; and our tubes were very 

 carefully filled with it by the kindness of F. 0. R. Gbtze, of 

 Leipzig. The preliminary work described in this paper was 

 chiefly done with a single tube containing gas at a pressure 

 of about one millimetre. The tube had a wide capillary and 

 was about 15 centim. in total length. In such a tube the red 

 glow of argon is readily obtained with a voltage of about two 

 thousand, but not with much less. A higher tension of gas 



