344 Dr. B. Hasselberg on the 



the simple induced current only was employed; but cases have 

 occurred, particularly with wider tubes, when the introduction 

 of a small air-break and Leyden jar has increased the distinct- 

 ness of the spectrum. This air-break must, however, be taken 

 so small that the characteristic lines do not appear hazy, since 

 in this case the second spectrum loses in brilliancy instead of 

 gaining. Four different tubes in all were employed for the 

 measurements. The first was a small tube containing hydro- 

 gen, about 3 inches long, of the usual form, whose capillary 

 was about 30 millim. long, and possessed an internal diameter 

 of about 0*5 millim. It was exhausted to a few tenths of a 

 millimetre. The second tube was a purchased hydrogen-tube. 

 The third tube had the well-known form introduced by Wtill- 

 ner ; but with the modification that a globe about an inch in 

 diameter, containing phosphoric anhydride, was attached to it 

 by grinding. The fourth tube, in order that it might be 

 placed longitudinally before the slit, was provided with a wide 

 cylinder at one end and at right angles to the capillary, which 

 was about 55 millim. long, as will be understood from the 

 figure (fig. 3, Plate X.). The internal diameters of the capil- 

 laries of these two tubes were respectively 1 and 2*5 millim., 

 and the pressure of the gas in the two cases 2 to 3 millim. 



It scarcely needs to be specially mentioned that in preparing 

 the gas used, as well as in filling the tubes, all the precautions 

 upon which its purity depends were strictly observed. The 

 gas evolved from zinc and sulphuric acid was therefore care- 

 fully freed from impurities by the use of known methods, and 

 introduced directly from the evolution-apparatus into the tube, 

 wdiich had been repeatedly heated and exhausted. Not until 

 the tubes had been in this way repeatedly washed out with 

 hydrogen were they closed at the pressure mentioned above, 

 and used for observation. The results of all the measure- 

 ments made of the spectrum are given in the following tables. 

 They contain the wave-lengths of all the lines visible under 

 the conditions employed, of which a large number, especially 

 in the green and blue portions of the spectrum, are so ex- 

 tremely weak that they approach the limits of visibility. It 

 might be thought superfluous to give here any more than the 

 final mean, but 1 have thought it desirable to give the separate 

 series of measurements, first, because the accuracy attained 

 can best be judged from them, and, secondly, because they 

 show the complete identity of the spectral phenomena with 

 the gas obtained in different ways. For the yellowish-red 

 part contained in Table A, which forms the characteristic 

 portion of the spectrum, the four series of measurements were 

 obtained by the use of the before-mentioned four tubes ; 



