Passage of Electricity through Gases. 163 



ends, be used. An examination of the number of discharges 

 through gas at different pressures led to the following con- 

 clusion : — 



III. The quantity of electricity necessary for effecting a dis- 

 charge in a spectrum-tube increases with the pressure, at first 

 rapidly, then more slowly until the pressure reaches 20 to 30 

 millims. Beyond this the quantity of electricity increases again 

 more rapidly, and becomes nearly proportional to the pressure. 



The reason for these results, which differ from those obtained 

 in the case of discharges between spheres in a free space, is to 

 be found in the charge on the glass envelope. With low 

 pressures the density of this is small, and diminishes only 

 slightly towards the capillary part of the tube, and the dis- 

 charge consequently proceeds from a larger part of the sur- 

 face of the electrodes. With higher pressures the density 

 of the charge on the glass, and its diminution towards the ca- 

 pillary part, are greater ; the discharge proceeds more from a 

 small surface near the ends of the electrodes ; cceteris paribus, 

 a smaller quantity of electricity suffices for the potential ne- 

 cessary to effect a discharge, while on the other hand the 

 higher pressure requires a larger quantity. The curve which 

 represents the relation between the quantity of electricity and 

 the pressure becomes more and more parallel to the line of 

 abscissae. The more the discharge is confined to the point 

 of the electrode, the more rapidly does the quantity of electri- 

 city increase with the pressure. 



By varying the length of the capillary part of the tube, the 

 author was led to the following conclusion : — 



IV. The quantity of electricity necessary to effect a discharge 

 is, within certain limits, independent of the length or breadth of 

 the tube joining the chambers which surround the electrodes. 



If these chambers are spherical, like the electrodes them- 

 selves, the quantity of electricity is not perceptibly dependent 

 upon the radius of the spheres, as may easily be concluded 

 from the absence of action of the electricity collected in the 

 glass chambers upon electrical points in their interior. Only 

 in the case of small glass spheres (in which the electricity dis- 

 tributed on the capillary part, on account of its less density 

 and smaller distance from the electrodes, obtains an influence) 

 is the quantity of electricity necessary for a discharge less. 

 Also, in spectrum-tubes the necessary quantity of electricity is 

 less with hydrogen than with air, &c. From this, if we neglect 

 secondary circumstances, we may conclude : — 



V. With a given gas at a given pressure the quantity of elec- 

 tricity necessary to effect a discharge depends only on the envi- 

 ronment of the electrodes. 



M2 



