18(5 Lord Armstrong. On a Multiple Induction 



[No. 22 shows a circular barrier of six wooden hemispheres, each 

 1J- inches diameter, and touching one another, formed round the centre 

 of the dust plate, and the dust was swept off the inner space before 

 making the experiment. The positive sparking point was level with 

 the top of the hemispheres, and the discharges were delivered from a 

 Leyden battery of two ^-gallon jars in series. It will be seen that 

 the whole of the dust outside the barrier is thrown into lines which 

 form arches over each touching point of the hemispheres, and that the 

 spandrils are filled up with inverted curves. Although it is not easy 

 to see how these arches and curves can be attributed to eddies, yet 

 appearances favour the view that the lines are due to the combined 

 effect of obstructed air drift and electric polarity. A barrier formed 

 by a continuous perpendicular screen of the same height as the hemi- 

 spheres almost entirely prevents the formation of lines. June 9.] 



Reverting to the subject of the development of heat at the nega- 

 tive side of the arc, the question arises, from what source is the heat 

 derived ? It cannot be acquired by conduction from the flame, for 

 mere conduction would heat both wires alike, nor can it be the result 

 of convection, for the arc is stagnant in the longitudinal direction. 

 The only explanation I can see is that the negative wire requires 

 time to take up the sudden gushes of current that come over from 

 the opposite side, and which, not being instantly carried onward, pro- 

 duce a tumultuous agitation of molecules at the receiving end of the 

 wire, resulting in the manifestation of heat. The thicker the wire 

 the greater the facility of passing on the current, and hence less heat 

 is evolved in a thick than in a thin wire. On the other hand, the 

 positive wire receives comparatively small returns, and therefore a 

 relatively thin wire suffices to pass on the pulsations without its being 

 heated to ignition. But the question remains, why is it that so much 

 less heat is produced by the spark discharge than by the arc dis- 

 charge ? Probably the chief reason is that the spark represents less 

 quantity of current, though higher in potential. Another reason may 

 be that the spark expends more energy in mechanical disturbance 

 than the arc. Then, again, there is the question, what is the relation 

 between the spark and the arc ? I cannot find that the spark is 

 possessed of mechanical impulse in the direction of its length any 

 more than the arc. Both produce lateral but not longitudinal dis- 

 persion. I have discharged a quick succession of powerful sparks in 

 a downward direction so as to pierce a piece of thick cardboard sus- 

 pended from a delicate balance-beam, bat without effecting any de- 

 cided disturbance of the balance. Every spark left a burr on each 

 side of the card apparently equal in size, which alone is sufficient 

 to show that the spark does not pierce like a needle. Probably 

 the arc and the spark are of much the same nature, the spark 



