380 PROCEEDINGS OF THE AMERICAN ACADEMY 



in the former case it shows itself chiefly as the flame-spark, in the 

 latter, as the thread-spark. 



The remarkable difference of tension according to the distance be- 

 tween the terminals, or, in other words, to the length of the spark, is 

 no doubt one, and perhaps the only, reason for the effects observed in 

 attempting to charge a Leyden jar by the inductive machine. When 

 we connect the outer coating of the jar with one of the terminals, and 

 bring the knob quite near the other, we see the broad flame-sparks 

 passing in quick succession ; but only a feeble charge is imparted, 

 however we may prolong the experiment. When, however, the knob 

 is held at a much greater distance from the terminal, a few of the long 

 thread-sparks are sufficient to charge the surface strongly. 



The beautiful phenomena of electrical light in rarefied gases, as ex- 

 hibited in the electrical egg and Gassiot's and Geissler's vacuum-tubes, 

 affoi'd many interesting subjects of inquiry. As the color of the light 

 is dependent on the specific nature of the gas, and as this is reduced to 

 an extreme degree of rarefaction, we have a means in some cases of 

 identifying such substances when their quantity is so minute as to defy 

 all other means of detection. With tubes of slender bore, affording, as 

 has been seen, a light of great intensity, we may obtain a brilliant pris- 

 matic spectrum, which, as Pliicker has shown, is marked in each case 

 by some characteristic peculiarity ; and with the same arrangement we 

 are able to trace the chemical changes which the enclosed gas or vapor 

 undergoes while subjected to the electrical action. 



Perhaps the most important observations in this connection are those 

 recently made by Gassiot, whose ingenious application of the absorbent 

 power of potassa has enabled him to approximate more nearly to an 

 absolute vacuum than any previous experimenter. In a tube thus pre- 

 pared, he has found that the gas may be so excessively rarefied as to 

 be unable to transmit the current, at this stage ceasing to be luminous. 

 We may therefore conclude that the old notion of a vacuum being a 

 good conductor, which was founded on the electric illumination of the 

 Torricellian space, is entirely erroneous, and that in all cases conduc- 

 tion is dependent on the presence of some form of ponderable matter. 



Adverting to the new evidences which these and other recent ex- 

 periments afforded of the electrical character of the Aurora, Professor 

 Kogers called attention to the action of a magnet on the electric light, 

 and more particularly to its power of arranging the illumination in me- 



