36 PROCEEDINGS OF THE AMERICAN ACADEMY 



a coiitiuuoiis current through the whole apparatus, iucludiug the 

 Sprengel pump which was connected with one of the openings. The 

 hydrogen was prepared from pure zinc and sulphuric acid, and most 

 carefully dried. "We would allow the gas to slowly pass through the 

 apparatus for twenty-four hours, then exhaust, and after exhaustion 

 heat the tube as hot as practicable under the circumstances, then pass 

 dry hydrogen and repeat the process several times. Notwithstanding 

 these precautions, we found, after a great many trials with different 

 tubes, that the fluted and more or less diffused spectrum always 

 accompanied the four principal hydrogen lines. It being then impossi- 

 ble to eliminate the diffused spectrum, we next tried alloying the 

 platinum electrodes with arsenic, and experimented with these in a 

 rarefied atmosphere of hydrogen, both with continuous discharge of 

 Rumkorff coil, and with interrupted discharge with Leyden jar. We 

 now obtained very definite arsenic bands, apparently the same iu both 

 cases ; but the effect was very momentary, and gave no opportunity 

 for measurement. The spectrum while it lasted was very striking ; 

 but, as soon as the arsenic upon the extreme point of the electrode 

 passed off, the characteristic spectrum disappeared. 



We were by this experience led to contrive the following apparatus, 

 by which we obtained the desired result, and the same may be useful 

 in experiments on the spectra of similar volatile substances. A longi- 

 tudinal section of the tube, one half the original size, is shown in Fig. 2 

 of plate. The portions AA' and A" are of rather coarse thermometer 

 tubing. J]I^ ' is a tube left open at J^, and drawn to a capillary 

 point at JB '. The substance to be examined, after being reduced to a 

 powder, is introduced through the opening at J^ until the tube is 

 about half full. Then one end of a platinum wire is buried in the 

 substance, and the other end is fused into the tube at J^, thus closing 

 the opening. After the hydrogen has been allowed to flow through 

 the tube a sufhcient length of time, the opening at A is closed by a 

 nipper-tap, and the tube is exhausted at A ". Now upon connecting 

 IB with the negative electrode, and C with the positive electrode, of a 

 small induction coil, we have the vapor of the substance in the tube 

 Uli' carried in the current through the tube A' where the spectrum 

 may be observed. 



One advantage of this particular form of tube is, that, in order to 

 compare the spectrum of our substance with that of hydrogen, we 

 have only to reverse the current, making C the negative pole, and then 

 all the lines except those of hydrogen at once disappear. 



The arsenic spectrum thus obtained is very brilliant, and consists of 



