under the Influence of the Electric Discharge. 425 



Of all the molecules present in the gas, there are always some 

 in which the atoms perform such vibrations that a very small 

 impulse is sufficient to separate them. Such impulses are 

 communicated to the atoms by the discharge in abundance ; 

 hence the hydrogen-lines appear already in tubes of relatively 

 great width, with a very small supply of electricity to each 

 molecule. So, on the other hand, there are molecules present 

 for the separation of whose atoms comparatively powerful im- 

 pulses are necessary ; the complete disappearance of the band 

 spectrum takes place therefore only very gradually. In exactly 

 the same way we understand why, with the discharges of the 

 induction-coil, we have the line spectrum and the band spec- 

 trum present together : it is to be observed that the discharge 

 of an induction-coil always consists of partial discharges of 

 varying strength. 



It was shown on p. 372 that, when electrodes of different 

 forms are employed, the heat produced in a capillary tube is 

 the same, and moreover the number of discharges does not 

 alter much. This fact shows that in a capillary tube, if we 

 have to do with discontinuous discharges, we can never have 

 spectra corresponding to quantities of electricity less than a 

 certain minimum value. 



The spectra which result from the smallest quantity of 

 electricity transmitted in each discharge, and the smallest 

 quantities of energy given off, are obtained by placing the two 

 electrodes close to each other and exhausting till the spark- 

 discharge ceases and the silent discharge alone takes place. 

 A cylindrical portion of the gas of several millimetres dia- 

 meter becomes luminous, and is traversed by a large number 

 of discharges. The energy given off may indeed fall so far 

 that scarcely any light at all is to be seen. 



The only means which we have of obtaining spectra of small 

 energy of internal motion with the necessary brilliancy consists 

 in the employment of wide tubes. We may get rid of the in- 

 fluence of thickness by employing tubes with axial line of sight; 

 but 1 have never been able to observe any difference between 

 nitrogen seen in this way and nitrogen seen transversely. 

 Hittorf has found similar negative results with sodium-vapour. 



In order to compare the action of the induction-coil with 

 that of the electrophorus-machine in producing spectra, 1 ex- 

 cited a Stohrer's induction-coil of medium size by the current 

 from four Bunsen cells, and employed as current-break an 

 automatic tuning-fork giving 100 complete vibrations ; the 

 one pole of the coil was connected directly with the earth, and 



