1882.] On the Spectra of Carbon and its Compounds. 127 



variation in the brightness of the hydrocarbon series than of the 

 cyanogen, in fact, the presence of magnesia rather favours the steady 

 formation of cyanogen. When the hydrocarbon spectrum is strong 

 the brilliancy and number of the cyanogen groups that are visible 

 are undoubtedly increased, so that the one variety of vibrations seems to 

 affect the other. This is easily accounted for by the chemical inter- 

 action which takes place between acetylene, nitrogen, and hydrocyanic 

 acid. The hydrocarbon spectrum is brought out at once in the 

 magnesia crucibles by moistening one of the poles. All such actions 

 seem to show that hydrogen is essentially connected with the produc- 

 tion of this fluted spectrum just as nitrogen is with the cyanogen 

 series. 



Arc Discharge in Fluids. 



The De Meritens arc, taken in water, shows the hydrocarbon 

 spectrum alone ; no cyanogen bands can be seen by eye observation, 

 even when ammonia or nitrate of potash is added to the water. In 

 this case the observations are rendered uncertain from the great 

 intensity of the continuous spectrum. If glycerine is used instead of 

 water no cyanogen groups of lines can be recognised, but on adding a 

 little nitrobenzol the set of three lines (about 4380) peculiar to the 

 cyanogen spectrum appear, this being the only group which can be 

 detected by the eye on the continuous background. This result sup- 

 ports the observations on the varying intensity of this group in 

 different gaseous media, and seems to show that conditions can be 

 found where it is the most characteristic group of cyanogen. These 

 three lines are easily seen in the spectrum of the arc taken in carbonic 

 acid, although they disappear from the spectrum of the arc taken in air. 



Vacuum Tube Spectra. 



In our former experiments with vacuum tubes we did not use a capil- 

 lary glass tube, but preferred to examine the photographic spectrum 

 obtained from a short spark taken between platinum wires. Objec- 

 tion has been taken to this plan of working on the ground that, as the 

 capillary form of vacuum tube increases the brilliancy of the spectrum, 

 particular lines or groups of the spectrum which otherwise would be 

 missed, might be revealed in them, and such tubes ought, therefore, to 

 have been employed. In order to answer this objection we have pre- 

 pared and examined vacuum tubes of this kind containing benzol and 

 benzol with naphthalene in solution, using all the precautions to avoid 

 the presence of nitrogen formerly described, and have always found 

 such tubes free from any trace of the cyanogen spectrum. When such 

 tubes are, however, examined daily, the cyanogen bands often appear 

 after a time, and this can be traced in all such cases to a leak or crack 

 at the point where the platinum is sealed into the glass. No per- 



