338 Dr. B. Hasselberg on the 



o 



spectrum; and Salet concludes therefore that Berthelot's view 

 is shown to be correct. However natural this may appear at 

 first sight, there is yet an objection which may be made, and 

 which appears to me to essentially alter the state of the case, 

 and to deprive Salet's experiments of most of their force. 

 We know that if a gas is rendered luminous in the way 

 described the temperature is comparatively low, or at least 

 decidedly below the limits it may attain when the discharge 

 passes directly through the tube. Consequently the intensity 

 of the light and of the spectrum will be smaller, and, unless 

 the coatings are very strongly charged, may be even so small 

 that only the stronger lines of the spectrum (in this case the 

 principal lines H«, H^, H y ) are visible. In order further to 

 test the admissibility of this explanation of Salet's experiment, 

 I have made the two following experiments : — A hydrogen- 

 tube (purchased ready filled), which showed the second hy- 

 drogen-spectrum very vividly when a powerful current was 

 sent through it, was provided with tinfoil coatings and 

 excited in the manner described. The light emitted by the 

 tube, usually brilliant and nearly white, was now feebly red, 

 and showed in the spectrum, besides mere traces* of the 

 second spectrum in the green, only H a and H y of small 

 intensity, and H^ relatively bright and diffuse. The same 

 experiment was then repeated with another tube, of the form 

 shown in fig. 2 (Plate X.). This tube was closed at one end, 

 A, by a plate-glass disk ground air-tight and cemented, in 

 order that it might be placed before the slit in a longitudinal 

 position (i. e. in tl e direction of the axis of the collimator), 

 and was united to the air-pump by means of the tube B. 

 After being repeatedly filled with pure hydrogen and exhausted, 

 it was finally excited, in the same way as before, with the gas 

 at a few millimetres 3 pressure, and examined with the spectro- 

 scope. If the tube was placed in the ordinary way before the 

 slit the spectrum was exactly that described, whilst it was 

 only necessary to place the tube longitudinally in order to see 

 the second spectrum fully developed. It follows therefore 

 that the absence of this spectrum in Salet's experiments 

 affords no certain proof that it belongs to the eliminated 

 carbon compounds and not to the hydrogen ; but their expla- 

 nation may just as well, and even with greater probability, be 

 sought in the diminution in intensity of light in the whole 

 phenomena due to the experimental conditions employed, a 

 diminution by which the already feeble lines of the second 

 spectrum must be first affected. 



* The lines X = 5015, 4930, 4G35. 



