342 Prof. J. Trowbridge on the Spectra of 



nected to the terminals of the battery, a slight heating was 

 sufficient to start a discharge and to cause the occluded 

 hydrogen to appear in the spectrum-tube. 



The operation of heating the spectrum-tubes is highly 

 important. Moreover, the discharge from a condenser 

 should be employed in the process of exhaustion, and the 

 operation should be continued for several hours*. 



Experience in the preparation of #-ray tubes is almost 

 essential to one who essays to investigate the spectra of gases. 

 The vacuum-pump I employed was a modification of Rapp's 

 pump. It was automatic. There was but one valve, and the 

 number of ground-glass joints with mercury seals was reduced 

 to a minimum. The hydrogen was produced by the electro- 

 lysis of phosphoric acid and doubly-distilled water. The gas- 

 holder contained many litres of the gas. The nitrogen was 

 prepared bypassing air through ammonia water, then through 

 heated copper gauze. The resulting gas was then passed 

 through tubes filled with caustic potash, and finally through 

 tubes of phosphoric pentoxide. 



The spectroscope consisted of a Rowland grating with a 

 ruled surface five inches by two inches, and a radius of 

 curvature of sixty inches. The dispersion was small on 

 account of the small number of lines to the inch ; but it had 

 the advantage of exhibiting the salient features of the spectra 

 on one plate with a dispersion of the order used by astro- 

 physicists, and thus enabled comparisons to be readily made. 

 The distance between the great II lines in the second spectrum 

 was approximately 2 millim. The camera was on an arm 

 pivoted at half the distance between the grating and the slit. 

 This arrangement was adopted for compactness. The grating 

 and camera were enclosed in a light-tight box ; and a lever 

 enabled me to expose the plate when this box was closed. 



Various liquid resistances were employed to regulate the 

 strength of the electric current. I finally used distilled water 

 which was kept running through a glass tube. My object 

 was not to measure the currents, but rather to keep them 

 within safe limits. I did not find it necessary therefore to 

 employ cadmium terminals in a solution of iodide of cadmium. 

 The direct battery-current gave me the white type of hydrogen 

 spectrum, while the second type was given by condenser-dis- 

 charges. A revolving drum provided with suitable metallic 

 strips enabled me to safely charge and discharge the con- 

 densers through the spectrum- tubes. 



* In this subject of exhaustion of tubes I am much indebted to sug- 

 gestions of Dr. William Rollins and Mr. Heinze of Boston, who have had 

 lung' experience in the preparation of r-ray tubes. 



