the Electric Discharge in highly Rarefied Gaseous Media. 525 



pagated through highly rarefied elastic fluids, may furnish some 

 new ideas as to the physical constitution of these bodies, and as 

 to the manner in which electricity is transmitted by them. 



§ 3. Supplement to the preceding Researches. 



The experiments described in §§ 1 and 2 had been already in 

 great part published in the Memoires de la Societe de Physique 

 et d J Histoire Naturelle de Geneve (vol. xvii. part 1), when I again 

 took up the subject and obtained results which have not hitherto 

 been published, and which appear to me to possess some interest 

 as throwing new light upon the nature of these phenomena. 



I began by taking two jars, exactly like the one with which I 

 had made my first experiments and which I have " described 

 above (p. 518), and I placed them so that the discharge of the 

 RuhmkorfFs coil could either be divided between them or be 

 made to traverse them in succession. In the former case, even 

 when care was taken to reduce the air in each as nearly as pos- 

 sible to the same degree of rarefaction, it was very difficult to 

 get the discharge to divide itself equally between them, at least 

 for any length of time. It would cease, after a short time, passing 

 through one of them, and pass wholly through the other ; if then 

 the passage of the discharge was interrupted for a moment and 

 then reestablished, it would often happen that the jar which 

 had previously transmitted the discharge would transmit it no 

 longer, and vice versa. These alternations are very probably 

 due to the difficulty that there is in producing, and in after- 

 wards maintaining, perfect identity between the two gaseous 

 media, notwitstanding all the pains bestowed upon this matter 

 at the beginning, differences of temperature and of molecular 

 arrangement occurring very quickly. 



When the discharge traverses the two jars in succession in- 

 stead of dividing itself between them, the rotation takes place 

 with equal rapidity in each, provided that care has been taken 

 to rarefy the air in both to exactly the same extent. A slight 

 difference is apparent only at very low pressures, in consequence 

 of the imperfect identity being more sensible at these pressures. 

 For instance : — 



Pressure. 



Number of turns in 15 seconds in 

 both jars alike. 



Ring positive. 



Ring negative. 



millims. 

 12 

 10 



8 



6 



14 



18 

 22 

 29 



12 

 14 

 16 

 21 



