THE PROPAGATION OF ELECTRICITY- 177 



not be betrayed, even by a galvanometer. But it does not the less exist, as M. 

 Gassiot has shown in operating with a pile of Grove at high tension, which, 

 with the same electrodes, and in the same medium, will give rise, first to strati- 

 fications, and afterwards, when the current has become continuous, to a voltaic 

 arch. 



The mechanical action of the series of discharges on the rarefied elastic fluid 

 may, indeed, be directly verified by the very marked oscillations of the column 

 of mercury of the manometer placed in communication with the clastic fluid, 

 which accompany the propagation of electricity iu that fluid. These oscilla- 

 tions rise to two or three tenths of a millimetre in hydrogen, under a pressure 

 of 16"^™. They begin to be sensible when once the stream passes, that is to 

 say, at 36°^™ of pressure ; attain their maximum of three-tenths of a millimetre 

 between 20 and 12™™ of pressure ; and diminish rapidly in descending from 12 

 to 5™™, at which last pressure they no longer take place. With nitrogen and at- 

 mospheric air, and employing the same tube 16 centimetres in length and 5 in 

 diameter, the oscillations begin, at the moment when the stream passes, under 

 the pressure of about 20™™; attain their maximum of from four to five tenths of 

 a millimetre between 12 and 8™™ of pressure ; and then continue to diminish 

 until 2 or 5™™, at which pressure they cease to be sensible. 



With the tube one metre in length, and even with that of 50 centimetres, I 

 have not succeeded in observing any appearance of oscillation accompanying 

 the transmission of the electric current, whatever might be the gas enclosed iu 

 these tubes, and whatever the pressure to which it was subjected. On the other 

 hand, 1 have obtained very distinct ones, of one and two tenths of a millimetre, 

 under pressures varying from 30 to 15™™, in a jar 20 centimetres in height by 

 16 in diameter, filled with rarefied hydrogen, and in which the electric stream 

 passed from a central ball to a ring 12 centimetres in diameter concentric to 

 that ball. This last result shows that the absence of oscillations in the long 

 lubes has less connexion with the volume of the gaseous medium, which is less 

 than in the vessel of the last experiment than with the influence of the sides 

 of the tubes which embarrass the movement of the gas. It is also a proof that 

 the oscillations proceed from a mechanical action, and not from an elevation of 

 temperature. As regards their intensity, the oscillations evidently depend on 

 the greater or less resistance which the gaseous medium opposes to the trans- 

 mission of the electric current, since the oscillations are more considerable with 

 nitrogen than with hydrogen, and diminish as the pressure does, reckoning from 

 a certain point of the pressure, which is that at which the discharge can take 

 place in a complete manner, and at which the intensity of the oscillations attains 

 its maximum. 



The stratification of electric light would appear then to be a phenomenon 

 analogous to the production of undulations of sound, that is to say, a mechanical 

 phenomenon proceeding from a succession of isochronous impulses communi- 

 cated to the rarefied gaseous column by the series of electric discharges rapidly 

 succeeding each other. We find a new proof favorable to this view of the phe- 

 nomenon in the perturbation which a displacement of the gaseous matter occa- 

 sions in the stratifications, and, consequently, in the disposition of the elastic 

 fluid Avhich permits those stratifications to appear. To produce this perturba- 

 tion, it suffices to introduce into the tube in which there is a rarefied elastic 

 fluid, fuul while the electricity is in process of propagation, an additional 

 quantity of the same gas already enclosed therein, so as to increase the pressure 

 by one-fourth or one-half of a millimetre at most. Let us see what then occurs 

 with hydrogen, remarking that the effects are the same with the three tubes, 

 respectively, 15, 50, and 100 centimetres in length. 



We begin by rarefying the gas to the extent of 2™™, so as to have the phe- 

 nomenon of the stratifications as distinct as possible. We then introduce a 

 small quantity of hydrogen ; if the introduction takes place on the side of the 

 12 S 



